Abstract

A general methodology to optimize the weight of power transmission structures is presented in this article. This methodology is based on the simulated annealing algorithm defined by Kirkpatrick in the early ‘80s. This algorithm consists of a stochastic approach that allows to explore and analyze solutions that do not improve the objective function in order to develop a better exploration of the design region and to obtain the global optimum. The proposed algorithm allows to consider the discrete behavior of the sectional variables for each element and the continuous behavior of the general geometry variables. Thus, an optimization methodology that can deal with a mixed optimization problem and includes both continuum and discrete design variables is developed. In addition, it does not require to study all the possible design combinations defined by discrete design variables. The algorithm proposed usually requires to develop a large number of simulations (structural analysis in this case) in practical applications. Thus, the authors have developed first order Taylor expansions and the first order sensitivity analysis involved in order to reduce the CPU time required. Exterior penalty functions have been also included to deal with the design constraints. Thus, the general methodology proposed allows to optimize real power transmission structures in acceptable CPU time.

Abstract

Abstract

Urban underground public spaces are relatively enclosed, with high daily passenger flow and diverse directions. Enhancing evacuation capabilities through dynamic guidance systems is a crucial solution. Therefore, this paper proposes a double-layer computational framework based on meta-heuristic integration of a bi-objective evacuation path optimization method. Firstly, the outer layer computation explores the population by combining the genetic algorithm and simulated annealing. It uses temperature parameters to probabilistically accept worse solutions, enhancing global search capability and avoiding local optima. Secondly, the inner layer calculates the fitness value of each solution from the outer layer, considering both the shortest path and minimum risk objectives. The path optimization is carried out by A* algorithm on the basis of constructing the safety matrix by using breadth-first search algorithm, and the distance matrix by using Dijkstra's Algorithm. Finally, this method can search for the shortest safe path, effectively guiding people safely from the starting point to the exit. The experiments show that compared to traditional evacuation path planning methods, the proposed method significantly improves path optimization capabilities, quickly planning the shortest and safest evacuation route. It can provide guidance for fire safety and emergency plans for evacuations.