Stackelberg game approach to a bi-objective robust design optimization

== Abstract ==

Robust design has received a great deal of attention from quality researchers in recent years, and a number of optimization methodologies based on the dual response format have been proposed. The majority of existing bi-objective optimization models concentrate on the trade-offs between the process mean and variability functions without investigating the interactions between control factors and quality characteristics. The primary objective of this research is to integrate the Stackelberg leadership model into the robust design procedure and propose a Stackelberg game-based robust design (SGRD) method to determine appropriate control factor settings by minimizing the values of desired optimization targets based on an analysis of possible combinations of input and output quality parameters. Herein, first, a bi-objective robust design optimization problem is formulated as a dual response model using response surface methodology (RSM). Second, the proposed SGRD model is developed via decomposition into two leader-follower game models. Finally, the mean square error (MSE) criterion is applied to evaluate models, and select non-dominated solutions in various situations. Numerical examples are used to demonstrate that the proposed method provides significant solutions in cases containing unidentified priorities between the dual responses and undiscovered correlations among several inputs and outcomes. In addition, according to the case study analysis, the proposed method is more efficient than the conventional dual response approach when dealing with bi-objective robust design optimization problems.

'''Keywords''': Robust Parameter Design, Bi-objective Optimization, Dual Response Approach, Game Theory, Stackelberg Competition model

==1. Introduction==

The subject of quality management and improvement has received widespread and consistent attention from researchers in the last century. Many studies on literature have shown that a high-efficiency and cost-efficient system plays a vital role in realistic production processes. Taguchi <span id="cite-_Ref76149437"></span>[[#_Ref76149437|[1]]] introduced a procedure for experimental design, which is also known as robust design, and that improves the robustness of quality characteristics by minimizing the response to noise factors. The main objective of robust design is to determine the optimal design settings for a product that can minimize the process variability and target deviation. Based on Taguchi's philosophy, subsequent experimental design methods have been developed to identify controlled parameter settings and, thus, improve the robustness of the production process.

Conventional robust design methods can be preliminarily classified as sequential and simultaneous optimization models. Vining and Myers <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]] investigated the application of a dual response approach to minimize the process variability by assigning the process mean at target with defining the priority criterion to separately optimize the estimated functions for the process mean and the standard deviation response based on RSM. Following the prioritized dual response format, Del Castillo and Montgomery <span id='cite-_Ref76149733'></span>[[#_Ref76149733|[3]]] and Copeland and Nelson <span id='cite-_Ref76149745'></span>[[#_Ref76149745|[4]]] presented improved optimization techniques for the situations in which one objective is achieved in advance (i.e., the process bias or variance is customer-specified). Instead of assigning a specific value to the optimization objective, the bias-specified and variability-specified robust design methods reached the goal of robust design by setting upper bounds for the process mean and standard deviation, respectively <span id='cite-_Ref76149761'></span>[[#_Ref76149761|[5,6]]]. Priority-based methods reduce the complexity of problems by transforming multi-objective models to constrained single-objective models. However, existing prioritization schemes are limited because the majority of feasible solutions in the solution space are removed once the first prioritized objective has finished optimization. In this case, the reliability of efficient solutions in objective space is significantly decreased.

The MSE method proposed by Lin and Tu <span id='cite-_Ref76150123'></span>[[#_Ref76150123|[7]]], which is based on the relaxed zero-bias assumption, deals with the dual response problem in a simultaneous mode. In terms of the MSE model and its simultaneous optimization order, the concept of weighted sum model is used by Tang and Xu <span id='cite-_Ref76150372'></span>[[#_Ref76150372|[8]]], Koksoy and Doganaksoy <span id='cite-_Ref76150383'></span>[[#_Ref76150383|[9]]] and Ding et al. <span id='cite-_Ref76150396'></span>[[#_Ref76150396|[10]]] to investigate the trade-offs between process bias and variance on different weights, instead of the default equivalent weight terms used by Lin and Tu <span id='cite-_Ref76150123'></span>[[#_Ref76150123|[7]]]. To find non-dominated solutions and non-convex Pareto frontier, Shin and Cho <span id='cite-_Ref76150480'></span>[[#_Ref76150480|[11]]] introduced the lexicographic weighted Tchebycheff method based on the L-infinity norm. The weight-based simultaneous models show good accuracy and efficiency when used to search for the optimal solutions of a robust design problem. Nevertheless, it is difficult to identify the actual weight terms in simultaneous models by quality operators in practice.

The majority of existing robust design optimization methodologies focus on the trade-offs between process bias and variability. However, the interaction between control variables and quality characteristics has always been ignored. To this end, we incorporate the Stackelberg game theory philosophy into classical sequential optimization methods that can consider the impact of control factors on output performance. In the proposed SGRD model in this paper, one object is optimized in advance as the leader, given the constraint of another object (the follower)'s best response function. The leader's objective function may be adjusted according to the follower's best response function, which is variable due to the presence of various controlled factors. Therefore, it is worth discovering situations with permutations and combinations among controlled factors to analyze the correlations between input factors and quality characteristics. Doing so means that every possible interaction between the preferred objective and non-preferred target are taken into account. In addition, in real-world production procedures, quality engineers are often confronted with scenarios in which the importance between process characteristics may be undiscovered or specified by customers. The numerical examples in Section 4 include two cases; the first has unknown priority for quality characteristics, while the second is based on customer-specified bias-prioritized and variance-prioritized situations.

Section 2 presents a literature review and background information regarding the robust parameter design model and Stackelberg leader-follower model. A description of the proposed SGRD model is given in Section 3. Two numerical examples are discussed and drawn in Section 4. Section 5 includes the conclusion and discussion of this paper. [[#img-1|Figure 1]] shows an overview of the proposed bi-objective SGRD optimization procedure.

<div id='img-1'></div>
{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: auto;max-width: auto;"
|-
|style="padding:10px;"|  [[File:Draft Shin 703976401-image1.png|centre|550x550px|]]
|- style="text-align: center; font-size: 75%;"
| colspan="1" style="padding:10px;"| '''Figure 1'''. Optimization procedure of the proposed SGRD model
|}

==2. Literature review==

=== 2.1 Robust parameter design===

The robust parameter design introduced by Genichi Taguchi is an experimental design technique that investigates the interactions between noise factors and outputs <span id='cite-_Ref76150513'></span>[[#_Ref76150513|[12]]]. Taguchi proposed a two-step model using the signal theory that considered both mean and standard deviation of quality characteristics. The goal of this method is to minimize the effects of uncontrolled factors by finding the optimal settings for the controlled factors, process bias, and variance, more accurately. Despite the ongoing debate about the accuracy of the Taguchi method for modeling noise factors, the basic philosophy of robust design has generally been applied in the quality management area.

From the perspective of optimization order, the robust design process can be said to consist of simultaneous and sequential optimization procedures. In simultaneous optimization models, the weight term is considered a crucial criterion on which to evaluate the trade-off among objectives. Lin and Tu <span id='cite-_Ref76150123'></span>[[#_Ref76150123|[7]]] integrated the MSE model, which takes both process bias and variance into account simultaneously, to solve the large variance problem in <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]]. As illustrated by Lin and Tu <span id='cite-_Ref76150123'></span>[[#_Ref76150123|[7]]], the MSE model's estimation is based on the assumption that the two objectives have the same weight. To identify more flexible and appropriate weight terms, Cho et al. <span id='cite-_Ref76150626'></span>[[#_Ref76150626|[13]]], Tang and Xu <span id='cite-_Ref76150372'></span>[[#_Ref76150372|[8]]], Koksoy and Doganaksoy <span id='cite-_Ref76150383'></span>[[#_Ref76150383|[9]]] and Ding et al. <span id='cite-_Ref76150396'></span>[[#_Ref76150396|[10]]] suggested using the concept of weighted sum based on the MSE model. In terms of the L-infinity norm, Kim and Lin <span id='cite-_Ref76150673'></span>[[#_Ref76150673|[14]]] used fuzzy logic and membership functions to optimize process mean and variance simultaneously. Instead of the weighted sum model, Shin and Cho <span id='cite-_Ref76150480'></span>[[#_Ref76150480|[11]]] suggested a lexicographical weighted Tchebycheff method to assign weight terms and find non-dominated solutions for a bi-objective optimization problem.

A sequential optimization procedure is also known as a priority-based performance method. RSM with the priority criterion was used in the first attempt at integration into the dual-response method in <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]]. To achieve bi-objective optimization, Vining and Myers <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]] assigned the mean value of the dual-response model at a customer-specified target and individually minimized the performance variance. Del Castillo and Montgomery <span id='cite-_Ref76149733'></span>[[#_Ref76149733|[3]]] and Copeland and Nelson <span id='cite-_Ref76149745'></span>[[#_Ref76149745|[4]]] incorporated the generalized reduced gradient method and Nelder-Mead simplex method into the dual-response model to guarantee a locally optimal solution. Kim and Cho <span id='cite-_Ref76150750'></span>[[#_Ref76150750|[15]]] and Tang and Xu <span id='cite-_Ref76150372'></span>[[#_Ref76150372|[8]]] improved the dual-response model using the goal programming method with a prioritized model. Shin and Cho <span id='cite-_Ref76149761'></span>[[#_Ref76149761|[5, 6]]] introduced bias-specified and variability-specified robust design models using the e-constraint method for customer-specified bias and variance upper bounds. Combined with the bias-specified model, Steenackers and Guillaume <span id='cite-_Ref76150810'></span>[[#_Ref76150810|[16]]] used the MSE model to determine the optimal design points. Regarding the variance-preference case, Kovach et al. <span id='cite-_Ref76150823'></span>[[#_Ref76150823|[17]]] introduced a variance-prioritized multi-response robust design model that contains multiple quality characteristics via goal programming.

=== 2.2  The Stackelberg leadership model ===
Game theory is the study of mathematical models of strategic interaction among rational decision-makers <span id='cite-_Ref76150835'></span>[[#_Ref76150835|[18]]]. Modern Game Theory is considered to have begun with the 1944 book "Theory of Games and Economics Behavior" by John von Neumann and Oskar Morgenstern. Game Theory was later fleshed out by John Forbes Nash to mathematically represent strategic interactions among rational agents. It has been wildly used in all social science domains, and its applications have been extended to the majority of other fields associated with decision making.

The Stackelberg model is an economic strategy game described by Heinrich Freiherr von Stackelberg in 1934. This leadership game holds that, in a market, a leader firm and follower firm compete for quantity and take actions in sequence. Chen and Cruz <span id='cite-_Ref76150849'></span>[[#_Ref76150849|[19]]] and Simaan and Cruz <span id='cite-_Ref76150858'></span>[[#_Ref76150858|[20]]] extended the static two-person non-cooperative nonzero-sum Stackelberg game to a dynamic game format with an asymmetric information criterion. Shimizu and Aiyoshi <span id='cite-_Ref76150867'></span>[[#_Ref76150867|[21]]] introduced a general bi-level Stackelberg game format and proposed an interior penalty method (i.e., barrier method) to solve the Stackelberg game-based non-linear min-max optimization problem. The penalty method was used in a static linear non-convex Stackelberg game model by Anandalingam and White <span id='cite-_Ref76150876'></span>[[#_Ref76150876|[22]]]. In addition, Anandalingam and White <span id='cite-_Ref76150876'></span>[[#_Ref76150876|[22]]] introduced and integrated the follower's rational reaction set into the leader's objective function. However, Lai <span id='cite-_Ref76150902'></span>[[#_Ref76150902|[23]]] pointed out that the use of the penalty method may restrict the leader's power to make decisions and applied membership functions in extended hierarchical/multi-level Stackelberg game models. A fuzzy approach was combined with the membership function and used for multi-level programming in <span id='cite-_Ref76150919'></span>[[#_Ref76150919|[24-26]]]. To search the Stackelberg equilibrium in a multi-level optimization problem, Liu <span id='cite-_Ref76150935'></span>[[#_Ref76150935|[27]]], D'Amato et al. <span id='cite-_Ref76150943'></span>[[#_Ref76150943|[28]]], and Hong et al. <span id='cite-_Ref76150950'></span>[[#_Ref76150950|[29]]] discussed the application of a genetic algorithm. The Stackelberg leader-follower competition model has been extensively applied in the security defense and supply chain domains, as discussed in <span id='cite-_Ref76150957'></span>[[#_Ref76150957|[30]]].

==3. Models and methods==

=== 3.1 Extensive-form game model ===
Game theory is widely recognized as an analytical tool to exploit the interactions among several participants; it has been applied to many fields. In terms of the properties of the game, a game can be classified as: cooperative/non-cooperative, simultaneous/sequential, perfect information/imperfect information, symmetric/asymmetric, and zero-sum/non-zero-sum. In this paper, we concentrate mostly on a non-cooperative sequential game model with perfect information.

In a general sequential game, which is also known in game theory as an extensive-form game, players take their actions in a specified sequence. Usually, the player who moves first is the front player, and the player who moves later is the rear player. The front player can make his decision in advance according to the anticipation of the rear player's movement. Then, the rear player can move based on the actions of the front player. Similar to a general sequential game, in a Stackelberg game, the player who makes the first decision is called the game leader, and the player who moves after the leader is known as the game follower. A particular extensive-form game with infinite action spaces can be used to represent the Stackelberg competition model, which is illustrated in [[#img-2|Figure 2]]. The potential actions of each player are shown fanning out from the player's decision node. Players can select and make decisions from the action space, which may be constrained by boundaries.

<div id='img-2'></div>
{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: auto;max-width: auto;"
|-
|style="padding:10px;"| [[File:Draft_Shin_703976401-image2.png|centre|400px]]
|- style="text-align: center; font-size: 75%;"
| colspan="1" style="padding:10px;"| '''Figure 2'''. An extensive-form Stackelberg game with infinite action space
|}


In sequential games, a backward induction methodology is applied to search for the sub-game perfect Nash equilibrium, where the consideration order is reversed so that it flows from the follower to the leader. Concerning the Stackelberg leadership model, a first partial derivative method is applied to analyze the later player's best response functions located in the follower's action space.

=== 3.2 The Stackelberg equilibrium: a general formulation ===

We identify two players in a non-cooperative sequential Stackelberg game that separately control the optimization objectives and control factors. Player 1's objective is to minimize his performance function <math display="inline">{f}_{1}(x,y)</math> and considers only factor <math display="inline">x</math>. Correspondingly, the objective of player 2 is to minimize his performance function <math display="inline">{f}_{2}(x,y)</math> and considers only factor <math display="inline">y</math>. Both functions are based on the factors <math display="inline">x</math> and <math display="inline">y</math>, which means that the strategies of two players are interacted. A variable set <math display="inline">(x,y)\in {(S}_{x},\, {S}_{y})</math> refers to a pair of rational strategy played by the two players, where <math display="inline">{S}_{x}</math> and <math display="inline">{S}_{y}</math> are the rational strategies set of the two players, respectively. The game ends when the game leader has optimized his performance function, which also means that an equilibrium is found. The Stackelberg equilibrium point and the associated Stackelberg equilibrium solution are denoted as <math display="inline">\left( {x}^{\ast },{y}^{\ast }\right)</math>  and <math display="inline">\lbrace {f}_{1}\left( {x}^{\ast },{y}^{\ast }\right) ,{f}_{2}\left( {x}^{\ast },{y}^{\ast }\right) \rbrace</math> , seperately. If player 1 acts as the leader, and player 2 acts as the follower, based on the Stackelberg game rules, player 1 optimizes his objective <math display="inline">{f}_{1}(x,y)</math> first, and player 2 optimizes <math display="inline">{f}_{2}(x,y)</math> later. The Stackelberg equilibrium point <math display="inline">({x}^{\ast },{y}^{\ast })</math> can be found in the objective function of player 1 if:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>{f}_{1}\left( {x}^{\ast },{y}^{\ast }\right) = \mathrm{inf}_{x \in {S}_x} {f}_{1}(x,\tilde{y}),</math>
|}
|  style="text-align: right;width: 5px;text-align: right;white-space: nowrap;"|(1)
|}


where <math display="inline">{f}_{1}(x,\tilde{y})</math> represents the objective function of player 1 that considers the best response of player 2 with the related controlled factor <math display="inline">\tilde{y}</math>, which can be found in the objective function of player 2 by:

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" |  <math>\tilde{y}=\,\mathrm{min}_{y\in {S}_{y}}\,{f}_{2}(x,y)=\frac{\partial {f}_{2}(x,y)}{\partial y}.</math>
|}
|}

Accordingly, if player 2 is leader and player 1 is the follower, the Stackelberg equilibrium solution can be obtained in the objective function of player 2 if:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>{{f}_{2}\left( {x}^{\ast },{y}^{\ast }\right) =}\,\mathrm{inf}_{y\in {S}_{y}}\,{f}_{2}(\tilde{x},y),</math>
|}
|  style="text-align: right;width: 5px;text-align: right;white-space: nowrap;"|(2)
|}

where <math display="inline">{f}_{2}(\tilde{x},y)</math> denotes the objective function of player 2 that takes the best response of player 1 with the corresponding controlled factor <math display="inline">\tilde{x}</math> into account, which can be found in the objective function of player 1 by:

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" |  <math>{\tilde{x}=}\,\mathrm{min}_{x\in {S}_{x}}\,{f}_{1}(x,y)=\frac{\partial {f}_{1}(x,y)}{\partial x}.</math>
|}
|}

=== 3.3 Bi-objective robust design model ===

In a bi-objective robust design model, the optimization target is to minimize both the process bias <math display="inline">{(\hat{\mu }\left( \boldsymbol\mathrm{x}\right) -\tau )}^{2}</math> and the process variance <math display="inline">{\hat{\sigma }}^{2}(\boldsymbol\mathrm{x})</math>, where <math display="inline">\boldsymbol\mathrm{x}</math> is denoted as the vector of input factors. Then, a bi-objective robust design optimization problem can be integrated and formulated as follows:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>\mathrm{min} {\left[ {(\hat{\mu }\left( \boldsymbol\mathrm{x}\right) -\tau )}^{2},\, {\hat{\sigma }}^{2}(\boldsymbol\mathrm{x})\right] }^{T}, \qquad s.t. \boldsymbol\mathrm{x}\in X.</math>
|}
|  style="text-align: right;text-align: right;white-space: nowrap;"|(3)
|}

where <math display="inline">\tau</math> refers to the target value of the process mean. <math display="inline">X</math> represents the feasible solution set, where <math display="inline">{R}^{n}\supseteq X=</math><math>\left\{ \boldsymbol\mathrm{x}\in {R}^{n}:\, {g}_{m}\left( \boldsymbol\mathrm{x}\right) \leq 0,\forall m\right\}</math>, and <math display="inline">{g}_{m}\left( \boldsymbol\mathrm{x}\right)</math>  denotes the ''m''th constraint. One possible solution <math display="inline">{\boldsymbol\mathrm{x}}_{1}\in X</math> is claimed that can dominate another solution <math display="inline">{\boldsymbol\mathrm{x}}_{2}\in X</math>, if <math display="inline">{\left\{ \hat{\mu }\left( {\boldsymbol\mathrm{x}}_{1}\right) -\tau \right\} }^{2}\leq {\left\{ \hat{\mu }\left( {\boldsymbol\mathrm{x}}_{2}\right) -\tau \right\} }^{2}</math> and <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{1})\leq {\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{2})</math>. An efficient solution <math display="inline">{\boldsymbol\mathrm{x}}^{\ast }\in X</math> (with the corresponding outcomes <math display="inline">{\left\{ \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right\} }^{2}</math> and <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>) is called the non-dominated solution if it cannot be dominated by any other solution in the model; it can then be regarded as an efficient trade-off between the objectives.

The conventional RSM is used to model and analyze response functions for estimating the process mean <math display="inline">\hat{\mu }\left( \boldsymbol\mathrm{x}\right)</math>  and standard deviation <math display="inline">\hat{\sigma }\left( \boldsymbol\mathrm{x}\right)</math>, which are:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>\hat{\mu }\left( \boldsymbol\mathrm{x}\right) =\, {\hat{\alpha }}_{0}+\, {\boldsymbol\mathrm{x}}^{T}\boldsymbol\mathrm{a}+</math><math>{\boldsymbol\mathrm{x}}^{T}\boldsymbol\mathrm{Ax}</math>
|}
|  style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;"|(4)
|}

where

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" | <math>\boldsymbol\mathrm{x}=\, \left[ \begin{matrix}{x}_{1}\\\begin{matrix}{x}_{2}\\\vdots \end{matrix}\\{x}_{k}\end{matrix}\right]\quad</math>  , <math>\quad\boldsymbol\mathrm{a}=</math><math>\, \left[ \begin{matrix}{\hat{\alpha }}_{1}\\\begin{matrix}{\hat{\alpha }}_{2}\\\vdots \end{matrix}\\{\hat{\alpha }}_{k}\end{matrix}\right]\quad </math>  and <math>\quad \boldsymbol\mathrm{A}=</math><math>\, \left[ \begin{matrix}{\hat{\alpha }}_{11}&\begin{matrix}{\hat{\alpha }}_{12}/2&\ldots \end{matrix}&{\hat{\alpha }}_{1k}/2\\{\hat{\alpha }}_{12}/2&\begin{matrix}{\hat{\alpha }}_{22}&\ldots \end{matrix}&{\hat{\alpha }}_{2k}/2\\\begin{matrix}\vdots \\{\hat{\alpha }}_{1k}/2\end{matrix}&\begin{matrix}\begin{matrix}\vdots \\{\hat{\alpha }}_{2k}/2\end{matrix}&\begin{matrix}\ddots \\\ldots \end{matrix}\end{matrix}&\begin{matrix}\vdots \\{\hat{\alpha }}_{kk}\end{matrix}\end{matrix}\right]</math> 
|}
|}

and

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>\hat{\sigma }\left( \boldsymbol\mathrm{x}\right) =\, {\hat{\beta }}_{0}+\, {\boldsymbol\mathrm{x}}^{T}\boldsymbol\mathrm{b}+\, {\boldsymbol\mathrm{x}}^{T}\boldsymbol\mathrm{Bx}</math>
|}
|  style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;"|(5)
|}

where

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" | <math>\boldsymbol\mathrm{x}=\, \left[ \begin{matrix}{x}_{1}\\\begin{matrix}{x}_{2}\\\vdots \end{matrix}\\{x}_{k}\end{matrix}\right]\quad</math>  , <math>\quad\boldsymbol\mathrm{b}=</math><math>\, \left[ \begin{matrix}{\hat{\beta }}_{1}\\\begin{matrix}{\hat{\beta }}_{2}\\\vdots \end{matrix}\\{\hat{\beta }}_{k}\end{matrix}\right]\quad</math>  and <math>\quad\boldsymbol\mathrm{B}=</math><math>\, \left[ \begin{matrix}{\hat{\beta }}_{11}&\begin{matrix}{\hat{\beta }}_{12}/2&\ldots \end{matrix}&{\hat{\beta }}_{1k}/2\\{\hat{\beta }}_{12}/2&\begin{matrix}{\hat{\beta }}_{22}&\ldots \end{matrix}&{\hat{\beta }}_{2k}/2\\\begin{matrix}\vdots \\{\hat{\beta }}_{1k}/2\end{matrix}&\begin{matrix}\begin{matrix}\vdots \\{\hat{\beta }}_{2k}/2\end{matrix}&\begin{matrix}\ddots \\\ldots \end{matrix}\end{matrix}&\begin{matrix}\vdots \\{\hat{\beta }}_{kk}\end{matrix}\end{matrix}\right]</math>
|}
|}

In these two second-order equations, parameters <math>\bf a</math> and <math>\bf A</math> are the vector and matrix of regression coefficients for the estimated process mean; parameters <math>\bf b</math> and <math>\bf B</math> are the vector and matrix of regression coefficients for estimated process variance. The ordinary least square method is used to obtain the values of <math>\bf a</math> and <math>\bf b</math>.

For an intuitive comparison among the Stackelberg solutions based on various situations, an MSE model is introduced as an evaluation criterion to determine the non-dominated solution, which is formulated as:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="text-align: right;margin:auto;width: 100%;"
|-
| <math display="block">\mathrm{MSE}\, =\, {(\hat{\mu }\left( \boldsymbol\mathrm{x}\right) -\tau )}^{2}+\, {\hat{\sigma }}^{2}(\boldsymbol\mathrm{x}),</math>
|}
|  style="text-align: right;width: 5px;text-align: right;white-space: nowrap;"|(6)
|}

=== 3.4 Proposed SGRD model ===

The integration procedure of the proposed bi-objective SGRD optimization problem from a Stackelberg game model is shown in Fig. 3. The process bias and variance can be identified as the two game players of the proposed SGRD model. Then, the objective functions introduced in the Stackelberg leader-follower model, <math display="inline">{f}_{1}\left( x,y\right)</math>  and <math display="inline">{f}_{2}\left( x,y\right)</math> , can be replaced by the estimated functions of process bias <math display="inline">{(\hat{\mu }\left( \boldsymbol\mathrm{x}\right) -\tau )}^{2}</math> and variance <math display="inline">{\hat{\sigma }}^{2}(\boldsymbol\mathrm{x})</math>, respectively. The optimization objectives of the bi-objective optimization problem <math display="inline">\left\{ {(\hat{\mu }\left( \boldsymbol\mathrm{x}\right) -\tau )}^{2},\, {\hat{\sigma }}^{2}(\boldsymbol\mathrm{x})\right\}</math>  and the input factors <math display="inline">\boldsymbol\mathrm{x}</math> are separately controlled by the process bias and variance. We can define the input factors that are controlled by the process bias are denoted by <math display="inline">\, {\boldsymbol\mathrm{x}}_{\mu }</math>, and <math display="inline">\, {\boldsymbol\mathrm{x}}_{\mu }\subseteq \boldsymbol\mathrm{x}</math>. Correspondingly,''' '''as the complement of <math display="inline">{\boldsymbol\mathrm{x}}_{\mu }</math>, the input factors that are controlled by the process variance are denoted by <math display="inline">{\boldsymbol\mathrm{\, x}}_{\sigma }</math>,where <math display="inline">{\boldsymbol\mathrm{x}}_{\sigma }=</math><math>{{\boldsymbol\mathrm{x}}_{\mu }}^{c}\subseteq \boldsymbol\mathrm{x}</math>. Similar to the Stackelberg game model, the response functions of process bias and variance are depended on the controlled factors <math display="inline">{\boldsymbol\mathrm{x}}_{\mu }</math> and <math display="inline">{\boldsymbol\mathrm{x}}_{\sigma }</math>. Thus, the objective functions for the process bias and variance can be written as <math display="inline">{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}</math> and <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math>, respectively. Based on the Stackelberg game rules, the target of process bias is to minimize <math display="inline">{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}</math> and considers only factor <math display="inline">{\boldsymbol\mathrm{x}}_{\mu }</math>. Accordingly, the target of process variance is to minimize <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math> and considers only factor <math display="inline">{\boldsymbol\mathrm{x}}_{\sigma }</math>.

<div id='img-3'></div>
{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: auto;max-width: auto;"
|-
|style="padding:10px;"|  [[File:Draft_Shin_703976401-image3.png|centre|600x600px]] 
|- style="text-align: center; font-size: 75%;"
| colspan="1" style="padding:10px;"| '''Figure 3'''. Integration procedure from a Stackelberg game model
|}


The proposed SGRD model can be investigated via decomposition into the bias-leader model and variance-leader model according to the optimization sequence. If the process bias is optimized in advance, the bias-leader model of the proposed SGRD model can be applied, and Eq. (1) of the two-player Stackelberg game model can be transformed into the bias-leader model. The Stackelberg equilibrium solution <math display="inline">\lbrace {\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) -\tau \right) }^{2},{\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) \rbrace</math>  with the Stackelberg equilibrium point <math display="inline">({\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast })</math> can be determined by the process bias if:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>\mathrm{{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) -\tau )}^{2}=}\,\mathrm{inf}_{{\boldsymbol\mathrm{x}}_{\mu }\in X}\,{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }\right) -\tau )}^{2}</math>
|}
|  style="text-align: right;width: 5px;text-align: right;white-space: nowrap;"|(7)
|}

where <math display="inline">{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }\right) -\tau )}^{2}</math> is the response function of the process bias that considers the best response of the process variance with the related controlled factor <math display="inline">{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }</math>, which can be performed in the response function of the process variance by:

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" | <math>\mathrm{{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }=}\,\mathrm{min}_{{\boldsymbol\mathrm{x}}_{\sigma }\in X}\,{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })=\frac{\partial {\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })}{\partial {\boldsymbol\mathrm{x}}_{\sigma }}</math>
|}
|}

and where <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math> is the estimated regression function for the process variability, which is performed using RSM.

Correspondingly, when the process variability is regarded as the game leader, the variance-leader of the proposed SGRD model can be implemented, and Eq. (2) can be transformed into the variance-leader model. Hence, the Stackelberg equilibrium can be determined by the process variance if:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math>\mathrm{{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast })=}\,\mathrm{inf}_{{\boldsymbol\mathrm{x}}_{\sigma }\in X}\,{\hat{\sigma }}^{2}({\tilde{\boldsymbol\mathrm{x}}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math>
|}
|  style="text-align: right;width: 5px;text-align: right;white-space: nowrap;"|(8)
|}

where <math display="inline">{\hat{\sigma }}^{2}({\tilde{\boldsymbol\mathrm{x}}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math> represents the response function of the process variance that considers the best response of the process bias with the corresponding controlled factor <math display="inline">{\tilde{\boldsymbol\mathrm{x}}}_{\mu }</math>, which can be performed in the response function of process bias by:

{| class="formulaSCP" style="width: 100%; text-align: left;" 
|-
| 
{| style="text-align: center; margin:auto;width: 100%;" 
|-
| style="text-align: center;" | <math>\mathrm{{\tilde{\boldsymbol\mathrm{x}}}_{\mu }=}\,\mathrm{min}_{{\boldsymbol\mathrm{x}}_{\mu }\in X}\,{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}=\frac{\partial {(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}}{\partial {\boldsymbol\mathrm{x}}_{\mu }}</math>
|}
|}

and where <math display="inline">\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right)</math>  is the estimated regression function for the process mean that can be obtained using RSM.

For clarity, the algorithm of the proposed SGRD is presented in a stepwise fashion in [[#tab-1|Table 1]]. In this paper, we used the MATLAB software package to perform the proposed methodology and identify the Stackelberg equilibrium for each optimization objective.

<div class="center" style="font-size: 75%;">'''Table 1'''. Algorithm of the proposed SGRD model</div>

<div id='tab-1'></div>
{| class="wikitable" style="margin: 1em auto 0.1em auto;border-collapse: collapse;font-size:85%;width:75%;" 
|-style="text-align:left"
| style="padding-top:10px;"| '''Step 1. Input data and assign controlled factors'''

:Separate the input factors <math display="inline">\boldsymbol\mathrm{x}</math> into controlled factors by the process bias <math display="inline">\left( {\boldsymbol\mathrm{x}}_{\mu }\right)</math>  and variance <math display="inline">\left( {\boldsymbol\mathrm{x}}_{\sigma }\right)</math>.

'''Step 2. Obtain estimated functions of the quality characteristics'''

:Use RSM to estimate the response functions of the process mean <math display="inline">\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right)</math>  and standard deviation <math display="inline">\hat{\sigma }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right)</math>.

'''Step 3. Formulate a bi-objective robust design model'''

:Formulate the estimated functions of process bias <math display="inline">{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}</math> and variance <math display="inline">{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })</math> as a bi-objective optimization problem:

:<math>\mathrm{min}~{\left[ {(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2},\, {\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })\right] }^{T}.</math>

'''Step 4. Evaluate the bias-leader model of the proposed SGRD model'''

:Evaluate the bias-leader model: <math display="inline">\left\{ \begin{matrix}\mathrm{{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) -\tau )}^{2}=}\,\mathrm{inf}_{{\boldsymbol\mathrm{x}}_{\mu }\in X}\,{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }\right) -\tau )}^{2}\\\mathrm{{\tilde{\boldsymbol\mathrm{x}}}_{\sigma }=}\,\mathrm{min}_{{\boldsymbol\mathrm{x}}_{\sigma }\in X}\,{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })=\frac{\partial {\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })}{\partial {\boldsymbol\mathrm{x}}_{\sigma }}\end{matrix}\right.</math>

:Find the Stackelberg equilibrium solution.

'''Step 5. Determine the non-dominated solution for the bias-leader model'''

:Alter possible combinations of <math display="inline">{\boldsymbol\mathrm{x}}_{\sigma }</math> and repeat Step 4.

:Determine the non-dominated solution for the bias-leader model based on the MSE criterion, and go to Step 6.

'''Step 6. Evaluate the variance-leader model of the proposed SGRD model'''

:Evaluate the variance-leader model: <math display="inline">\left\{ \begin{matrix}\mathrm{{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast })=}\,\mathrm{inf}_{{\boldsymbol\mathrm{x}}_{\sigma }\in X}\,{\hat{\sigma }}^{2}({\tilde{\boldsymbol\mathrm{x}}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma })\\\mathrm{{\tilde{\boldsymbol\mathrm{x}}}_{\mu }=}\,\mathrm{min}_{{\boldsymbol\mathrm{x}}_{\mu }\in X}\,{(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}=\frac{\partial {(\hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu },{\boldsymbol\mathrm{x}}_{\sigma }\right) -\tau )}^{2}}{\partial {\boldsymbol\mathrm{x}}_{\mu }}\end{matrix}\right.</math>

:Find the Stackelberg equilibrium solution.

'''Step 7. Determine the non-dominated solution for the bias-leader model'''

:Alter possible combinations of <math display="inline">{\boldsymbol\mathrm{x}}_{\mu }</math> and repeat Step 6.

:Determine the non-dominated solution for the variance-leader model based on the MSE criterion, and go to Step 8.

'''Step 8. Determine the most efficient solution for the proposed SGRD model'''

:Compare the non-dominated solution of the bias-leader model with that of the variance-leader model based on the MSE criterion.

:Determine the most efficient solution <math display="inline">\left\{ {\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) -\tau \right) }^{2},{\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast }\right) \right\}</math>  of the proposed model with the corresponding controlled factor setting <math display="inline">({\boldsymbol\mathrm{x}}_{\mu }^{\ast },{\boldsymbol\mathrm{x}}_{\sigma }^{\ast })</math> based on the MSE criterion.
|}

==4. Numerical examples==

=== 4.1. Example 1 ===
A printing machine example is discussed in <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]]. The authors investigated how the printer's ability ( <math display="inline">y</math>) to apply colored inks to package labels is affected by speed ( <math display="inline">{x}_{1}</math>), pressure ( <math display="inline">{x}_{2}</math>), and distance ( <math display="inline">{x}_{3}</math>). This is a three-level factorial experimental design with three runs ( <math display="inline">{r}_{i}</math>) at each design point. Detailed experimental data is shown in Table 2, where <math display="inline">\overline{y}</math> and <math display="inline">s</math> represent the process mean and process standard deviation, respectively.<span id="cite-_Ref76149715"></span>
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 2.''' Printing study data [[#_Ref76149715|[2]]]
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |Run ( <math display="inline">{r}_{i}</math>)
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{1}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{3}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>\overline{y}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>s</math>
|-
| style="border-top: 1pt solid black;text-align: center;" |1
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |34
| style="border-top: 1pt solid black;text-align: center;" |10
| style="border-top: 1pt solid black;text-align: center;" |28
| style="border-top: 1pt solid black;text-align: center;" |24.00
| style="border-top: 1pt solid black;text-align: center;" |12.50
|-
| style="text-align: center;" |2
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |115
| style="text-align: center;" |116
| style="text-align: center;" |130
| style="text-align: center;" |120.30
| style="text-align: center;" |8.40
|-
| style="text-align: center;" |3
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |192
| style="text-align: center;" |186
| style="text-align: center;" |263
| style="text-align: center;" |213.70
| style="text-align: center;" |42.80
|-
| style="text-align: center;" |4
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |82
| style="text-align: center;" |88
| style="text-align: center;" |88
| style="text-align: center;" |86.00
| style="text-align: center;" |3.70
|-
| style="text-align: center;" |5
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |44
| style="text-align: center;" |178
| style="text-align: center;" |188
| style="text-align: center;" |136.70
| style="text-align: center;" |80.40
|-
| style="text-align: center;" |6
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |322
| style="text-align: center;" |350
| style="text-align: center;" |350
| style="text-align: center;" |340.70
| style="text-align: center;" |16.20
|-
| style="text-align: center;" |7
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |141
| style="text-align: center;" |110
| style="text-align: center;" |86
| style="text-align: center;" |112.30
| style="text-align: center;" |27.60
|-
| style="text-align: center;" |8
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |259
| style="text-align: center;" |251
| style="text-align: center;" |259
| style="text-align: center;" |256.30
| style="text-align: center;" |4.60
|-
| style="text-align: center;" |9
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |290
| style="text-align: center;" |280
| style="text-align: center;" |245
| style="text-align: center;" |271.70
| style="text-align: center;" |23.60
|-
| style="text-align: center;" |10
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |81
| style="text-align: center;" |81
| style="text-align: center;" |81
| style="text-align: center;" |81.00
| style="text-align: center;" |0.00
|-
| style="text-align: center;" |11
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |90
| style="text-align: center;" |122
| style="text-align: center;" |93
| style="text-align: center;" |101.70
| style="text-align: center;" |17.70
|-
| style="text-align: center;" |12
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |319
| style="text-align: center;" |376
| style="text-align: center;" |376
| style="text-align: center;" |3570
| style="text-align: center;" |32.90
|-
| style="text-align: center;" |13
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |180
| style="text-align: center;" |180
| style="text-align: center;" |154
| style="text-align: center;" |171.30
| style="text-align: center;" |15.00
|-
| style="text-align: center;" |14
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |372
| style="text-align: center;" |372
| style="text-align: center;" |372
| style="text-align: center;" |372.00
| style="text-align: center;" |0.00
|-
| style="text-align: center;" |15
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |541
| style="text-align: center;" |658
| style="text-align: center;" |396
| style="text-align: center;" |501.70
| style="text-align: center;" |92.50
|-
| style="text-align: center;" |16
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |288
| style="text-align: center;" |192
| style="text-align: center;" |312
| style="text-align: center;" |264.00
| style="text-align: center;" |63.50
|-
| style="text-align: center;" |17
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |432
| style="text-align: center;" |336
| style="text-align: center;" |513
| style="text-align: center;" |427.00
| style="text-align: center;" |88.60
|-
| style="text-align: center;" |18
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |713
| style="text-align: center;" |725
| style="text-align: center;" |754
| style="text-align: center;" |730.70
| style="text-align: center;" |21.10
|-
| style="text-align: center;" |19
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |364
| style="text-align: center;" |99
| style="text-align: center;" |199
| style="text-align: center;" |220.70
| style="text-align: center;" |133.80
|-
| style="text-align: center;" |20
| style="text-align: center;" |0
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |232
| style="text-align: center;" |221
| style="text-align: center;" |266
| style="text-align: center;" |239.70
| style="text-align: center;" |23.50
|-
| style="text-align: center;" |21
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |408
| style="text-align: center;" |415
| style="text-align: center;" |443
| style="text-align: center;" |422.00
| style="text-align: center;" |18.50
|-
| style="text-align: center;" |22
| style="text-align: center;" |-
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |182
| style="text-align: center;" |233
| style="text-align: center;" |182
| style="text-align: center;" |199.00
| style="text-align: center;" |29.40
|-
| style="text-align: center;" |23
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |507
| style="text-align: center;" |515
| style="text-align: center;" |434
| style="text-align: center;" |485.30
| style="text-align: center;" |44.60
|-
| style="text-align: center;" |24
| style="text-align: center;" |+
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |846
| style="text-align: center;" |535
| style="text-align: center;" |640
| style="text-align: center;" |673.70
| style="text-align: center;" |158.20
|-
| style="text-align: center;" |25
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |236
| style="text-align: center;" |126
| style="text-align: center;" |168
| style="text-align: center;" |176.70
| style="text-align: center;" |55.50
|-
| style="text-align: center;" |26
| style="text-align: center;" |0
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |660
| style="text-align: center;" |440
| style="text-align: center;" |403
| style="text-align: center;" |501.00
| style="text-align: center;" |138.90
|-
| style="border-bottom: 1pt solid black;text-align: center;" |27
| style="border-bottom: 1pt solid black;text-align: center;" |+
| style="border-bottom: 1pt solid black;text-align: center;" |+
| style="border-bottom: 1pt solid black;text-align: center;" |+
| style="border-bottom: 1pt solid black;text-align: center;" |878
| style="border-bottom: 1pt solid black;text-align: center;" |991
| style="border-bottom: 1pt solid black;text-align: center;" |1 161
| style="border-bottom: 1pt solid black;text-align: center;" |1 010.00
| style="border-bottom: 1pt solid black;text-align: center;" |142.50
|}

The fitted response surface functions for the process mean and standard deviation in <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]] are:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math display="inline">\hat{\mu }\left( \boldsymbol\mathrm{x}\right) =\, 327.6+\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{a}+</math><math>{\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{AX}</math>,
|}
|  style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;"|(9)
|}


where

<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
<math display="inline">\boldsymbol\mathrm{a}=\, \left[ \begin{matrix}177\\109.4\\131.5\end{matrix}\right] \quad</math> and <math display="inline">\quad \boldsymbol\mathrm{A}=</math><math>\, \left[ \begin{matrix}32&66&75.5\\66&-22.4&43.6\\75.5&43.6&-29.1\end{matrix}\right]</math>  ,</div>

and

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math display="inline">\hat{\sigma }\left( \boldsymbol\mathrm{x}\right) =\, 34.9+\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{b}+</math><math>\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{BX}</math>,
|}
|  style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;"|(10)
|}


where

<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
<math display="inline">\boldsymbol\mathrm{b}=\, \left[ \begin{matrix}11.5\\15.3\\29.2\end{matrix}\right] \quad</math> and <math display="inline">\quad \boldsymbol\mathrm{B}=</math><math>\, \left[ \begin{matrix}4.2&7.7&5.1\\7.7&-1.3&14.1\\5.1&14.1&16.8\end{matrix}\right]</math>  ,</div>

The feasible solution set <math display="inline">X=\left\{ \boldsymbol\mathrm{x}\in {R}^{3}:\, g\left( \boldsymbol\mathrm{x}\right) =\right. </math><math>\left. \left| {x}_{m}\right| -1\leq 0,\, m=1,2,3\right\} .</math> The target value of the process mean is assumed to be 500. The process mean and variance can be considered by “Nominal the best” and “Smaller the better” types of quality characteristics. Generally, it is possible to analyze this bi-objective optimization problem by decomposing the proposed SGRD model into a bias-leader model and a variance-leader model. In addition, we consider different correlations between input variables and quality characteristics to analyze all possible situations and find the follower's best responses in each model.

====4.1.1. Bias-leader model====

First, we consider a situation in which the process bias is optimized first as the leader in a Stackelberg game. In this situation, process variance is the follower. The detailed results using Eq. (7) of the bias-leader model are shown in Table 3.
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 3.''' Bias-leader Stackelberg game results (example 1)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |SGRD

Bias-leader model
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Variance control <math display="inline">{x}_{1}</math>
| style="border-top: 1pt solid black;text-align: center;" |-0.8262
| style="border-top: 1pt solid black;text-align: center;" |-0.5404
| style="border-top: 1pt solid black;text-align: center;" |-0.0783
| style="border-top: 1pt solid black;text-align: center;" |113 390
| style="border-top: 1pt solid black;text-align: center;" |475.21
| style="border-top: 1pt solid black;text-align: center;" |113 865.21
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{2}</math>
| style="text-align: center;" |0.1818
| style="text-align: center;" |1
| style="text-align: center;" |-1
| style="text-align: center;" |66 609
| style="text-align: center;" |630.14
| style="text-align: center;" |67 239.14
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{3}</math>
| style="text-align: center;" |1
| style="text-align: center;" |-0.0496
| style="text-align: center;" |-1
| style="text-align: center;" |42 477
| style="text-align: center;" |1 066.35
| style="text-align: center;" |43 543.35
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{2}</math>
| style="text-align: center;" |-1
| style="text-align: center;" |-0.0406
| style="text-align: center;" |-0.5465
| style="text-align: center;" |127 830
| style="text-align: center;" |378.26
| style="text-align: center;" |128 208.26
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{3}</math>
| style="text-align: center;" |-0.5029
| style="text-align: center;" |-0.5814
| style="text-align: center;" |-0.5487
| style="text-align: center;" |123 470
| style="text-align: center;" |325.32
| style="text-align: center;" |123 795.32
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{2},{x}_{3}</math>
| style="text-align: center;" |-0.0450
| style="text-align: center;" |0.3283
| style="text-align: center;" |-1
| style="text-align: center;" |101 940
| style="text-align: center;" |500.01
| style="text-align: center;" |102 440.01
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{2},{x}_{3}</math>
| style="border-bottom: 1pt solid black;text-align: center;" |-0.8046
| style="border-bottom: 1pt solid black;text-align: center;" |-0.1676
| style="border-bottom: 1pt solid black;text-align: center;" |-0.6766
| style="border-bottom: 1pt solid black;text-align: center;" |129 900
| style="border-bottom: 1pt solid black;text-align: center;" |365.31
| style="border-bottom: 1pt solid black;text-align: center;" |130 265.31
|}

As shown in the table, if we consider that the process variance is highly correlated to <math display="inline">{x}_{3}</math>, the minimum MSE value (43 543.35) with the related control factor setting <math display="inline">({x}_{1}^{\ast },\, {x}_{2}^{\ast },\, {x}_{3}^{\ast })=(1,\, -0.0496,\, -</math><math>1)</math> is obtained when <math display="inline">{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}=</math><math>42\, 477</math> and <math display="inline">{\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}^{\ast }\right) =</math><math>1\, 066.35</math>.

====4.1.2. Variance-leader model====

Next, we consider a situation in which the process variance is prioritized as the Stackelberg game leader. The results, which are derived using Eq. (8) of the variance-leader model, are presented in Table 4. Based on the assumption of process bias controls <math display="inline">{x}_{1}</math> and <math display="inline">{x}_{2}</math>, we obtain the efficient result that <math display="inline">MSE=</math><math>2\, 049.10</math>. The process variance is 2 049.10 in the zero-bias condition, and the corresponding variable setting is <math display="inline">({x}_{1}^{\ast },\, {x}_{2}^{\ast },\, {x}_{3}^{\ast })=(1,\, 0.3229,\, -</math><math>0.3911)</math>.
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 4.''' Variance-leader Stackelberg game results (example 1)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<span id="OLE_LINK7"></span><span id="OLE_LINK8"></span>SGRD

Variance-leader model
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Bias control <math display="inline">{x}_{1}</math>
| style="border-top: 1pt solid black;text-align: center;" |-1
| style="border-top: 1pt solid black;text-align: center;" |-1
| style="border-top: 1pt solid black;text-align: center;" |-0.6225
| style="border-top: 1pt solid black;text-align: center;" |161 760
| style="border-top: 1pt solid black;text-align: center;" |360.44
| style="border-top: 1pt solid black;text-align: center;" |162 120.44
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{2}</math>
| style="text-align: center;" |-0.6593
| style="text-align: center;" |0.5935
| style="text-align: center;" |-0.9013
| style="text-align: center;" |132 930
| style="text-align: center;" |308.73
| style="text-align: center;" |133 238.73
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{3}</math>
| style="text-align: center;" |-1
| style="text-align: center;" |-1
| style="text-align: center;" |0.2131
| style="text-align: center;" |145 830
| style="text-align: center;" |466.27
| style="text-align: center;" |146 296.27
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{2}</math>
| style="text-align: center;" |1
| style="text-align: center;" |0.3229
| style="text-align: center;" |-0.3911
| style="text-align: center;" |0
| style="text-align: center;" |2 049.10
| style="text-align: center;" |2 049.10
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{3}</math>
| style="text-align: center;" |1
| style="text-align: center;" |-0.1030
| style="text-align: center;" |-0.0892
| style="text-align: center;" |0
| style="text-align: center;" |2 063.07
| style="text-align: center;" |2 063.07
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{2},{x}_{3}</math>
| style="text-align: center;" |1
| style="text-align: center;" |-0.0820
| style="text-align: center;" |-0.1069
| style="text-align: center;" |0
| style="text-align: center;" |2 057.15
| style="text-align: center;" |2 057.15
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{2},{x}_{3}</math>
| style="border-bottom: 1pt solid black;text-align: center;" |0.9991
| style="border-bottom: 1pt solid black;text-align: center;" |-0.1369
| style="border-bottom: 1pt solid black;text-align: center;" |-0.0591
| style="border-bottom: 1pt solid black;text-align: center;" |0
| style="border-bottom: 1pt solid black;text-align: center;" |2 075.53
| style="border-bottom: 1pt solid black;text-align: center;" |2 075.53
|}

====4.1.3. Results and Discussion====
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 5.''' Comparison of the optimization results (example 1)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<span id="OLE_LINK1"></span><span id="OLE_LINK2"></span>Method
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Dual response approach
| style="border-top: 1pt solid black;text-align: center;" |0.614
| style="border-top: 1pt solid black;text-align: center;" |0.228
| style="border-top: 1pt solid black;text-align: center;" |0.100
| style="border-top: 1pt solid black;text-align: center;" |0
| style="border-top: 1pt solid black;text-align: center;" |2 680.96
| style="border-top: 1pt solid black;text-align: center;" |2 680.96
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Proposed SGRD model
| style="border-bottom: 1pt solid black;text-align: center;" |1
| style="border-bottom: 1pt solid black;text-align: center;" |0.3229
| style="border-bottom: 1pt solid black;text-align: center;" |-0.3911
| style="border-bottom: 1pt solid black;text-align: center;" |0
| style="border-bottom: 1pt solid black;text-align: center;" |2 049.10
| style="border-bottom: 1pt solid black;text-align: center;" |2 049.10
|}

Comparing the non-dominated results of the bias-leader and variance-leader models in example 1 based on the MSE criterion shows that the variance-leader model provides a more efficient outcome, and it is more appropriate to optimize process variance in advance when the process bias is treated as a restriction. In addition, <math display="inline">{x}_{1}</math> and <math display="inline">{x}_{2}</math> are regarded as the controlled factors for process bias, and the solution set is <math display="inline">\left\{ {\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2},\, {\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}^{\ast }\right) \right\} =</math><math>(0,\, 2\, 049.10)</math>.

A comparison of the significant results of example 1 between the dual response approach in <span id='cite-_Ref76149715'></span>[[#_Ref76149715|[2]]] and the proposed SGRD model is exhibited in Table 5. As evidenced by the results, the proposed SGRD model is more efficient than the dual response model based on the same zero-bias condition. Since the dual response approach is a pure priority-based optimization method, the comparison also shows the importance of considering the correlations between controlled factors and optimization objectives. The contour plots for the process mean and standard deviation are shown in Fig. 4, where the Stackelberg equilibrium solutions of the bias-leader model and variance-leader model are denoted by dots. The solution obtained by using dual response approach is plotted as a star, and the most efficient solution provided by the proposed SGRD model is marked as a circle.
[[File:Draft_Shin_703976401-image4.png|centre|thumb|600x600px|'''Fig. 4.''' Contour plots for process mean and standard deviation (example 1)]]

=== 4.2. Example 2 ===
The experimental data in example 2 concerns the bare silicon wafer production process of an injection-molding company for a large motor corporation. The dataset appears in <span id='cite-_Ref76151131'></span>[[#_Ref76151131|[6]]]. Three essential factors <math display="inline">({x}_{i})</math>, mold temperature, injection flow rate, and cooling rate, are considered to affect the coating thickness ( <math display="inline">y</math>) of bare silicon wafers. The coded and uncoded units of control factors are shown in Table 6, and the spherical central composite design is presented in Table 7.<span id="cite-_Ref76151131"></span>
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 6.''' Uncoded and coded units of three control factors [[#_Ref76151131|[6]]]
|-
! rowspan="2" style="border-top: 1pt solid black;text-align: center;" |<span id="OLE_LINK5"></span><span id="OLE_LINK6"></span>Levels of

coded variables
! colspan="3" style="border-top: 1pt solid black;text-align: center;" |Uncoded or original variables <math display="inline">({x}_{i})</math>
|-
! style="border-bottom: 1pt solid black;text-align: center;" |Mold temperature (F)

<math>{x}_{1}</math>
! style="border-bottom: 1pt solid black;text-align: center;" |Injection flow rate (lbs/sec)

<math>{x}_{2}</math>
! style="border-bottom: 1pt solid black;text-align: center;" |Cooling rate (F/sec)

<math>{x}_{3}</math>
|-
| style="border-top: 1pt solid black;text-align: center;" |<math>-\alpha \, (-1.682)</math>
| style="border-top: 1pt solid black;text-align: center;" |99.6
| style="border-top: 1pt solid black;text-align: center;" |40
| style="border-top: 1pt solid black;text-align: center;" |4.8
|-
| style="text-align: center;" |-1
| style="text-align: center;" |140.5
| style="text-align: center;" |46
| style="text-align: center;" |15
|-
| style="text-align: center;" |0
| style="text-align: center;" |200.5
| style="text-align: center;" |55
| style="text-align: center;" |30
|-
| style="text-align: center;" |+1
| style="text-align: center;" |260.5
| style="text-align: center;" |64
| style="text-align: center;" |45
|-
| style="border-bottom: 1pt solid black;text-align: center;" |<math>+\alpha \, (+1.682)</math>
| style="border-bottom: 1pt solid black;text-align: center;" |301.4
| style="border-bottom: 1pt solid black;text-align: center;" |70
| style="border-bottom: 1pt solid black;text-align: center;" |55.2
|}

<nowiki> </nowiki><span id="cite-_Ref76151131"></span> 
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 7.''' The layout of central composite design [[#_Ref76151131|[6]]]
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{1}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{3}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{y}_{4}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>\overline{y}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>s</math>
|-
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |-
| style="border-top: 1pt solid black;text-align: center;" |76.30
| style="border-top: 1pt solid black;text-align: center;" |80.50
| style="border-top: 1pt solid black;text-align: center;" |77.70
| style="border-top: 1pt solid black;text-align: center;" |81.10
| style="border-top: 1pt solid black;text-align: center;" |78.90
| style="border-top: 1pt solid black;text-align: center;" |2.28
|-
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |79.10
| style="text-align: center;" |81.20
| style="text-align: center;" |78.80
| style="text-align: center;" |79.60
| style="text-align: center;" |79.68
| style="text-align: center;" |1.07
|-
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |82.50
| style="text-align: center;" |81.50
| style="text-align: center;" |79.50
| style="text-align: center;" |80.90
| style="text-align: center;" |81.10
| style="text-align: center;" |1.25
|-
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |72.30
| style="text-align: center;" |74.30
| style="text-align: center;" |75.70
| style="text-align: center;" |72.70
| style="text-align: center;" |73.75
| style="text-align: center;" |1.56
|-
| style="text-align: center;" |-
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |70.60
| style="text-align: center;" |72.70
| style="text-align: center;" |69.90
| style="text-align: center;" |71.50
| style="text-align: center;" |71.18
| style="text-align: center;" |1.21
|-
| style="text-align: center;" |+
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |74.10
| style="text-align: center;" |77.90
| style="text-align: center;" |76.20
| style="text-align: center;" |77.10
| style="text-align: center;" |76.33
| style="text-align: center;" |1.64
|-
| style="text-align: center;" |-
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |78.50
| style="text-align: center;" |80.00
| style="text-align: center;" |76.20
| style="text-align: center;" |75.30
| style="text-align: center;" |77.50
| style="text-align: center;" |2.14
|-
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |+
| style="text-align: center;" |84.90
| style="text-align: center;" |83.10
| style="text-align: center;" |83.90
| style="text-align: center;" |83.50
| style="text-align: center;" |83.85
| style="text-align: center;" |0.77
|-
| style="text-align: center;" |-1.682
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |74.10
| style="text-align: center;" |71.80
| style="text-align: center;" |72.50
| style="text-align: center;" |71.90
| style="text-align: center;" |72.58
| style="text-align: center;" |1.06
|-
| style="text-align: center;" |1.682
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |76.40
| style="text-align: center;" |78.70
| style="text-align: center;" |79.20
| style="text-align: center;" |79.30
| style="text-align: center;" |78.40
| style="text-align: center;" |1.36
|-
| style="text-align: center;" |0
| style="text-align: center;" |-1.682
| style="text-align: center;" |0
| style="text-align: center;" |79.20
| style="text-align: center;" |80.70
| style="text-align: center;" |81.00
| style="text-align: center;" |82.30
| style="text-align: center;" |80.80
| style="text-align: center;" |1.27
|-
| style="text-align: center;" |0
| style="text-align: center;" |1.682
| style="text-align: center;" |0
| style="text-align: center;" |77.90
| style="text-align: center;" |76.40
| style="text-align: center;" |76.90
| style="text-align: center;" |77.40
| style="text-align: center;" |77.15
| style="text-align: center;" |0.65
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |-1.682
| style="text-align: center;" |82.40
| style="text-align: center;" |82.70
| style="text-align: center;" |82.60
| style="text-align: center;" |83.10
| style="text-align: center;" |82.70
| style="text-align: center;" |0.29
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |1.682
| style="text-align: center;" |79.70
| style="text-align: center;" |82.40
| style="text-align: center;" |81.00
| style="text-align: center;" |81.20
| style="text-align: center;" |81.08
| style="text-align: center;" |1.11
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |70.40
| style="text-align: center;" |70.60
| style="text-align: center;" |70.80
| style="text-align: center;" |71.10
| style="text-align: center;" |70.73
| style="text-align: center;" |0.30
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |70.90
| style="text-align: center;" |69.70
| style="text-align: center;" |69.00
| style="text-align: center;" |69.90
| style="text-align: center;" |69.88
| style="text-align: center;" |0.78
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |70.70
| style="text-align: center;" |71.90
| style="text-align: center;" |71.70
| style="text-align: center;" |71.20
| style="text-align: center;" |71.38
| style="text-align: center;" |0.54
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |70.20
| style="text-align: center;" |71.00
| style="text-align: center;" |71.50
| style="text-align: center;" |70.40
| style="text-align: center;" |70.78
| style="text-align: center;" |0.59
|-
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |0
| style="text-align: center;" |71.50
| style="text-align: center;" |71.10
| style="text-align: center;" |71.20
| style="text-align: center;" |70.00
| style="text-align: center;" |70.95
| style="text-align: center;" |0.66
|-
| style="border-bottom: 1pt solid black;text-align: center;" |0
| style="border-bottom: 1pt solid black;text-align: center;" |0
| style="border-bottom: 1pt solid black;text-align: center;" |0
| style="border-bottom: 1pt solid black;text-align: center;" |71.00
| style="border-bottom: 1pt solid black;text-align: center;" |70.40
| style="border-bottom: 1pt solid black;text-align: center;" |70.90
| style="border-bottom: 1pt solid black;text-align: center;" |69.90
| style="border-bottom: 1pt solid black;text-align: center;" |70.55
| style="border-bottom: 1pt solid black;text-align: center;" |0.51
|}

The estimated functions for the process mean and standard deviation in <span id="cite-_Ref76151131"></span>[[#_Ref76151131|[6]]] derived using RSM are:

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;"  
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math display="inline">\hat{\mu }\left( \boldsymbol\mathrm{x}\right) =\, 72.21+\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{a}+</math><math>{\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{AX}</math>,
|}
| style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;" |(11)
|}

where

<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
<math display="inline">\boldsymbol\mathrm{a}=\, \left[ \begin{matrix}0.59\\-0.35\\-0.01\end{matrix}\right] \quad</math> and <math display="inline">\quad \boldsymbol\mathrm{A}=</math><math>\, \left[ \begin{matrix}0.28&0.09&1.66\\0.09&1.29&1.51\\1.66&1.51&1.85\end{matrix}\right]</math>  ,</div>

and

{| class="formulaSCP" style="width: 100%;border-collapse: collapse;width: 100%;text-align: center;" 
|-
| 
{| style="vertical-align: top;margin:auto;width: 100%;"
|-
| <math display="inline">\hat{\sigma }\left( \boldsymbol\mathrm{x}\right) =\, 2.55+\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{b}+</math><math>\, {\boldsymbol\mathrm{X}}^{T}\boldsymbol\mathrm{BX}</math>,
|}
|  style="text-align: right;vertical-align: top;width: 5px;text-align: right;white-space: nowrap;"|(12)
|}


where

<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
<math display="inline">\boldsymbol\mathrm{b}=\, \left[ \begin{matrix}0.38\\-0.43\\0.56\end{matrix}\right] \quad</math> and <math display="inline">\quad \boldsymbol\mathrm{B}=</math><math>\, \left[ \begin{matrix}0.49&-0.47&0.72\\-0.47&0.61&-0.12\\0.72&-0.12&0.85\end{matrix}\right]</math>  ,</div>

The feasible solution set <math display="inline">X=\left\{ \boldsymbol\mathrm{x}\in {R}^{3}:\, g\left( \boldsymbol\mathrm{x}\right) =\right. </math><math>\left. \sum _{m=1}^{3}{x}_{m}^{2}-3\leq 0,\, m=1,2,3\right\} .</math> The target coating thickness is specified by the customer and is set at 71.14. The goal of the bi-objective optimization problem is to minimize process bias and variability. Based on the same optimization progress in example 1, the proposed SGRD model can be decomposed into a bias-leader model and variance-leader model, which are stated below.

====4.2.1. Bias-leader model====

In the bias-leader model, process variance is imagined as the optimization constraints while optimizing the process bias, and various combinations of the controlled variables have been taken into account. Eq. (7) of the proposed SGRD model is applied, and the experimental outcome is shown in Table 8.
  
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 8.''' Bias-leader Stackelberg game results (example 2)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |SGRD

Bias-leader model
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Variance control <math display="inline">{x}_{1}</math>
| style="border-top: 1pt solid black;text-align: center;" |-1.3436
| style="border-top: 1pt solid black;text-align: center;" |-0.5222
| style="border-top: 1pt solid black;text-align: center;" |0.9602
| style="border-top: 1pt solid black;text-align: center;" |0.0316
| style="border-top: 1pt solid black;text-align: center;" |11.8185
| style="border-top: 1pt solid black;text-align: center;" |11.8501
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{2}</math>
| style="text-align: center;" |-1.4786
| style="text-align: center;" |-0.1643
| style="text-align: center;" |0.5381
| style="text-align: center;" |0
| style="text-align: center;" |9.1059
| style="text-align: center;" |9.1059
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{3}</math>
| style="text-align: center;" |-1.6885
| style="text-align: center;" |0.0014
| style="text-align: center;" |0.3858
| style="text-align: center;" |0.0039
| style="text-align: center;" |10.1086
| style="text-align: center;" |10.1125
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{2}</math>
| style="text-align: center;" |-0.8403
| style="text-align: center;" |0.0954
| style="text-align: center;" |0.6783
| style="text-align: center;" |0.5461
| style="text-align: center;" |8.5949
| style="text-align: center;" |9.1410
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{3}</math>
| style="text-align: center;" |-0.0778
| style="text-align: center;" |0.2155
| style="text-align: center;" |-0.2813
| style="text-align: center;" |1.2160
| style="text-align: center;" |5.7586
| style="text-align: center;" |6.9746
|-
| style="text-align: center;" |Variance control <math display="inline">{x}_{2},{x}_{3}</math>
| style="text-align: center;" |-1.6743
| style="text-align: center;" |-0.2570
| style="text-align: center;" |0.3616
| style="text-align: center;" |0.0303
| style="text-align: center;" |9.7644
| style="text-align: center;" |9.7947
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Variance control <math display="inline">{x}_{1},{x}_{2},{x}_{3}</math>
| style="border-bottom: 1pt solid black;text-align: center;" |-0.0146
| style="border-bottom: 1pt solid black;text-align: center;" |0.3173
| style="border-bottom: 1pt solid black;text-align: center;" |-0.3008
| style="border-bottom: 1pt solid black;text-align: center;" |1.2398
| style="border-bottom: 1pt solid black;text-align: center;" |5.7351
| style="border-bottom: 1pt solid black;text-align: center;" |6.9749
|}

According to the MSE criterion, the solution set <math display="inline">\left\{ {\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2},\, {\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}^{\ast }\right) \right\} =</math><math>\left( 1.2160,\, 5.7586\right)</math>  with input variable setting <math display="inline">({x}_{1}^{\ast },\, {x}_{2}^{\ast },\, {x}_{3}^{\ast })=(-0.0778,\, 0.2155,\, -</math><math>0.2813)</math> has been identified as the non-dominated result in this model, where the MSE value is 6.9746. Moreover, controlled factors <math display="inline">{x}_{1}</math> and <math display="inline">{x}_{3}</math> are assumed to be significantly correlated with the game follower, which is process variance.

====4.2.2. Variance-leader model====

Using Eq. (8) from the proposed SGRD model yields the results listed in Table 9. The non-dominated solution set of the variance-leader model is <math display="inline">\left\{ {\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2},\, {\hat{\sigma }}^{2}\left( {\boldsymbol\mathrm{x}}^{\ast }\right) \right\} =</math><math>(1.2414,\, 5.7351),</math> with a minimum MSE value of 6.9765. The corresponding control variable setting is <math display="inline">({x}_{1}^{\ast },\, {x}_{2}^{\ast },\, {x}_{3}^{\ast })=(-0.0158,\, 0.3130,\, -</math><math>0.3021)</math>, and the input factor <math display="inline">{x}_{2}</math> is regarded as controlled by the process bias.
  
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 9.''' Variance-leader Stackelberg game results (example 2)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |SGRD

Variance-leader model
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Bias control <math display="inline">{x}_{1}</math>
| style="border-top: 1pt solid black;text-align: center;" |-0.0056
| style="border-top: 1pt solid black;text-align: center;" |0.3090
| style="border-top: 1pt solid black;text-align: center;" |-0.3703
| style="border-top: 1pt solid black;text-align: center;" |1.3680
| style="border-top: 1pt solid black;text-align: center;" |5.7528
| style="border-top: 1pt solid black;text-align: center;" |7.1208
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{2}</math>
| style="text-align: center;" |-0.0158
| style="text-align: center;" |0.3130
| style="text-align: center;" |-0.3021
| style="text-align: center;" |1.2414
| style="text-align: center;" |5.7351
| style="text-align: center;" |6.9765
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{3}</math>
| style="text-align: center;" |0.0623
| style="text-align: center;" |0.4282
| style="text-align: center;" |-0.2000
| style="text-align: center;" |1.2609
| style="text-align: center;" |5.8279
| style="text-align: center;" |7.0888
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{2}</math>
| style="text-align: center;" |0.0124
| style="text-align: center;" |0.3570
| style="text-align: center;" |-0.3789
| style="text-align: center;" |1.3797
| style="text-align: center;" |5.7581
| style="text-align: center;" |7.1378
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{3}</math>
| style="text-align: center;" |0.5313
| style="text-align: center;" |0.8449
| style="text-align: center;" |-0.5805
| style="text-align: center;" |2.2640
| style="text-align: center;" |6.5000
| style="text-align: center;" |8.7640
|-
| style="text-align: center;" |Bias control <math display="inline">{x}_{2},{x}_{3}</math>
| style="text-align: center;" |0.0945
| style="text-align: center;" |0.2044
| style="text-align: center;" |-0.1231
| style="text-align: center;" |1.1757
| style="text-align: center;" |6.0388
| style="text-align: center;" |7.2145
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Bias control <math display="inline">{x}_{1},{x}_{2},{x}_{3}</math>
| style="border-bottom: 1pt solid black;text-align: center;" |1.3620
| style="border-bottom: 1pt solid black;text-align: center;" |0.5836
| style="border-bottom: 1pt solid black;text-align: center;" |-0.8465
| style="border-bottom: 1pt solid black;text-align: center;" |1.8876
| style="border-bottom: 1pt solid black;text-align: center;" |8.5498
| style="border-bottom: 1pt solid black;text-align: center;" |10.4374
|}

====4.2.3. Results and Discussion====
  
{| style="width: 100%;border-collapse: collapse;"  
|+'''Table 10.''' Comparison of the optimization results (example 2)
|-
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |Method
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{1}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{2}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{x}_{3}^{\ast }</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\left( \hat{\mu }\left( {\boldsymbol\mathrm{x}}^{\ast }\right) -\tau \right) }^{2}</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |<math>{\hat{\sigma }}^{2}({\boldsymbol\mathrm{x}}^{\ast })</math>
! style="border-top: 1pt solid black;border-bottom: 1pt solid black;text-align: center;" |MSE
|-
| style="border-top: 1pt solid black;text-align: center;" |Dual response approach
| style="border-top: 1pt solid black;text-align: center;" |-1.4561
| style="border-top: 1pt solid black;text-align: center;" |-0.1456
| style="border-top: 1pt solid black;text-align: center;" |0.5596
| style="border-top: 1pt solid black;text-align: center;" |0
| style="border-top: 1pt solid black;text-align: center;" |9.0854
| style="border-top: 1pt solid black;text-align: center;" |9.0854
|-
| style="border-bottom: 1pt solid black;text-align: center;" |Proposed SGRD model
| style="border-bottom: 1pt solid black;text-align: center;" |-0.0778
| style="border-bottom: 1pt solid black;text-align: center;" |0.2155
| style="border-bottom: 1pt solid black;text-align: center;" |-0.2813
| style="border-bottom: 1pt solid black;text-align: center;" |1.2160
| style="border-bottom: 1pt solid black;text-align: center;" |5.7586
| style="border-bottom: 1pt solid black;text-align: center;" |6.9746
|}

A comparison of the non-dominated results of the bias-leader and variance-leader models in example 2 based on the MSE criterion shows that the result of the bias-leader model is slightly better than that of the variance-leader model.

The bias-leader model of the proposed SGRD method is next compared with the dual response approach; the results are presented in Table 10. Based on the zero-bias assumption, the process variance and MSE values obtained using the dual response approach are equal to 9.0854. With the more relaxed bias assumption, the proposed SGRD model provides a more efficient solution. Fig. 5 is the contour plots for the process mean and standard deviation, where the notation dot, star, and circle represent the Stackelberg equilibrium solution, the solution obtained by using dual response approach, and the most efficient solution provided by the proposed SGRD model, respectively.
[[File:Draft_Shin_703976401-image5.png|centre|thumb|600x600px|'''Fig. 5.''' Contour plots for the process mean and standard deviation (example 2)]]

==5. Conclusion==

Conventional priority-based robust design optimization methodologies are based on the assumption of preset priorities. However, in a real-world production process, the exact priorities and weights among several objectives are always unknown. In order to find a reasonable solution based on that situation, in this paper, we proposed a bi-objective robust design optimization method based on game theory. Furthermore, existing priority-based approaches place an emphasis on investigating the trade-offs among several different optimization objectives, but it is worth considering the interaction relationships among input control factors and quality characteristics. To this end, we introduced the proposed SGRD optimization procedure to integrate the effects of control factors. As shown in the numerical examples, the proposed SGRD model can be used to analyze all scenarios with different optimization sequences and various combinations of control factors. Therefore, it is helpful for quality engineers to determine the most appropriate condition from the perspective of game theory. In this paper, we discussed a bi-objective optimization model based on a single quality response, which can be extended to multi-response robust design optimization models in the future.

== Acknowledgment ==
This work was supported by the National Research Foundation of Korea (NRF) (No. NRF- 2019R1F1A1060067).

==References==

<span id='_Ref76149437'></span>[[#cite-_Ref76149437|[1]]] Taguchi, G. Introduction to Quality Engineering. Asian Productivity Organization. UNIPUB, White Plains, NY, 1986.

<span id='_Ref76149715'></span>[[#cite-_Ref76149715|[2]]] Vining, G. G., Myers, R. H. Combining Taguchi and response surface philosophies: a dual response approach. Journal of quality technology, 22(1):38-45, 1990.

<span id='_Ref76149733'></span>[[#cite-_Ref76149733|[3]]] Del Castillo, E., Montgomery, D. C. A nonlinear programming solution to the dual response problem. Journal of Quality Technology, 25(3):199-204, 1993.

<span id='_Ref76149745'></span>[[#cite-_Ref76149745|[4]]] Copeland, K. A., Nelson, P. R. Dual response optimization via direct function minimization. Journal of Quality Technology, 28(3):331-336, 1996.

<span id='_Ref76149761'></span>[[#cite-_Ref76149761|[5]]] Shin, S., Cho, B. R. Bias-specified robust design optimization and its analytical solutions. Computers & Industrial Engineering, 48(1):129-140, 2005.

<span id='_Ref76151131'></span>[[#cite-_Ref76151131|[6]]] Shin, S., Cho, B. R. Robust design models for customer-specified bounds on process parameters. Journal of Systems Science and Systems Engineering, 15(1):2-18, 2006.

<span id='_Ref76150123'></span>[[#cite-_Ref76150123|[7]]] Lin, D. K., Tu, W. Dual response surface optimization. Journal of quality technology, 27(1):34-39, 1995.

<span id='_Ref76150372'></span>[[#cite-_Ref76150372|[8]]] Tang, L. C., Xu, K. A unified approach for dual response surface optimization. Journal of quality technology, 34(4):437-447, 2002.

<span id='_Ref76150383'></span>[[#cite-_Ref76150383|[9]]] Köksoy, O., Doganaksoy, N. Joint optimization of mean and standard deviation using response surface methods. Journal of Quality Technology, 35(3):239-252, 2003.

<span id='_Ref76150396'></span>[[#cite-_Ref76150396|[10]]] Ding, R., Lin, D. K., Wei, D. Dual-response surface optimization: a weighted MSE approach. Quality engineering, 16(3):377-385, 2004.

<span id='_Ref76150480'></span>[[#cite-_Ref76150480|[11]]] Shin, S., Cho, B. R. Studies on a biobjective robust design optimization problem. IIE Transactions, 41(11):957-968, 2009.

<span id='_Ref76150513'></span>[[#cite-_Ref76150513|[12]]] Brewer, K., Carraway, L., Ingram, D. Forward Selection as a Candidate for Constructing Nonregular Robust Parameter Designs. Technical Report. Arkansas State University, 2010.

<span id='_Ref76150626'></span>[[#cite-_Ref76150626|[13]]] Cho, B. R., Philips, M. D., Kapur, K. C. Quality improvement by RSM modeling for robust design. The 5th Industrial Engineering Research Conference, pp. 650-655, May 1996.

<span id='_Ref76150673'></span>[[#cite-_Ref76150673|[14]]] Kim, K. J., Lin, D. K. Dual response surface optimization: a fuzzy modeling approach. Journal of Quality Technology, 30(1):1-10, 1998.

<span id='_Ref76150750'></span>[[#cite-_Ref76150750|[15]]] Kim, Y. J., Cho, B. R. Development of priority-based robust design. Quality Engineering, 14(3):355-363, 2002.

<span id='_Ref76150810'></span>[[#cite-_Ref76150810|[16]]] Steenackers, G., Guillaume, P. Bias-specified robust design optimization: A generalized mean squared error approach. Computers & Industrial Engineering, 54(2):259-268, 2008.

<span id='_Ref76150823'></span>[[#cite-_Ref76150823|[17]]] Kovach, J., Cho, B. R., Antony, J. Development of a variance prioritized multiresponse robust design framework for quality improvement. The International Journal of Quality & Reliability Management, 26(4):380, 2009.

<span id='_Ref76150835'></span>[[#cite-_Ref76150835|[18]]] Myerson, R. B. Game theory. Harvard university press, 1997.

<span id='_Ref76150849'></span>[[#cite-_Ref76150849|[19]]] Chen, C. L., Cruz, J. Stackelburg solution for two-person games with biased information patterns. IEEE Transactions on Automatic Control, 17(6):791-798, 1972.

<span id='_Ref76150858'></span>[[#cite-_Ref76150858|[20]]] Simaan, M., Cruz, J. B. On the Stackelberg strategy in nonzero-sum games. Journal of Optimization Theory and Applications, 11(5):533-555, 1973.

<span id='_Ref76150867'></span>[[#cite-_Ref76150867|[21]]] Shimizu, K., Aiyoshi, E. A new computational method for Stackelberg and min-max problems by use of a penalty method. IEEE Transactions on Automatic Control, 26(2):460-466, 1981.

<span id='_Ref76150876'></span>[[#cite-_Ref76150876|[22]]] Anandalingam, G., White, D. J. A solution method for the linear static Stackelberg problem using penalty functions. IEEE Transactions on automatic control, 35(10):1170-1173, 1990.

<span id='_Ref76150902'></span>[[#cite-_Ref76150902|[23]]] Lai, Y. J. Hierarchical optimization: a satisfactory solution. Fuzzy sets and systems, 77(3):321-335, 1996.

<span id='_Ref76150919'></span>[[#cite-_Ref76150919|[24]]] Shih, H. S., Lai, Y. J., Lee, E. S. Fuzzy approach for multi-level programming problems. Computers & Operations Research, 23(1):73-91, 1996.

[25] Farahi, M. H., Ansari, E. A new approach to solve Multi-objective linear bilevel programming problems. Journal of Mathematics and Computer Science, 1(4):313-320, 2010.

[26] Roy, S. K., Maiti, S. K. Stochastic bi level programming with multi-choice for Stackelberg game via fuzzy programming. International Journal of Operational Research, 29(4):508-530, 2017.

<span id='_Ref76150935'></span>[[#cite-_Ref76150935|[27]]] Liu, B. Stackelberg-Nash equilibrium for multilevel programming with multiple followers using genetic algorithms. Computers & Mathematics with Applications, 36(7):79-89, 1998.

<span id='_Ref76150943'></span>[[#cite-_Ref76150943|[28]]] D'Amato, E., Daniele, E., Mallozzi, L., Petrone, G., Tancredi, S. A hierarchical multimodal hybrid Stackelberg–Nash GA for a leader with multiple followers game. Dynamics of Information Systems: Mathematical Foundations. Springer, New York, NY, pp. 267-280, 2012.

<span id='_Ref76150950'></span>[[#cite-_Ref76150950|[29]]] Hong, Z., Chu, C., Zhang, L. L., Yu, Y. Optimizing an emission trading scheme for local governments: A Stackelberg game model and hybrid algorithm. International Journal of Production Economics, 193:172-182, 2017.

<span id='_Ref76150957'></span>[[#cite-_Ref76150957|[30]]] He, X., Prasad, A., Sethi, S. P., Gutierrez, G. J. A survey of Stackelberg differential game models in supply and marketing channels. Journal of Systems Science and Systems Engineering, 16(4):385-413, 2007.