Evolutionary algorithm for optimization regarding the planning of topological facilities in layout of a shipyard

Abstract

The purpose of this study is to contribute the approach to the problem of optimization regarding the planning of topological facilities in layout of a shipyard, with the objective of finding a robust solution to the problem by improving the solution space search through refining the genetic operators. For this, the computational results of the evolutionary algorithm proposed by Choi with changes made by the authors, being: 1) the use of the Partially-Matched Crossover (PMX) genetic operator; 2) the use of a recursive expression in the topological optimization step in addition to implementing the Biased Random-Key Genetic Algorithm (BRKGA) for the purpose of comparing the results. As a plan of the computational experiments two groups of experiments were performed: 1) with the parameters and variables of the work of Choi, in order to validate the efficiency and effectiveness of the AE proposed in this work and; 2) with the parameters and variables of the work of Choi with Department 03 fixed in the position of the best solution found in the 1st group of experiments (position 11 of the topological Grid). Each group contains 50 experiments with 100 iterations and variation of the number of individuals from 100 to 80,000 individuals. As a result, a better solution characterized by the reduction of material handling costs, of 11,816 presented by Choi, for 11,489 monetary units of cost, found from the changes made by the authors of the original proposal of the evolutionary algorithm and the use of BRKGA.