Título: An Adaptive Pursuit Genetic Algorithm for Solving Job-Shop Scheduling Problems
Autores: Ferreira, Guilherme;Bernardino, Heder
When a metaheuristic is used for solving Job-Shop Scheduling Problems (JSPs), ones need to select the correct movement operators and theirs parameters to improve the results for them. However, the correct setup for a problem is a hard work and it is problem-dependent. In this work, we propose the use of an Adaptive Genetic Algorithm (AGA) to automatically control the techniques contained in its framework, while it is solving the problem. An Adaptive Pursuit Method with Extreme Credit Assignment is used to select the movement operators (crossover and mutation techniques) and its parameters, and select the Local Search rate. The algorithm is tested in instances provided by a well-known generator for JSPs. The results show superior performance and reliability when compared with a standard genetic algorithm.
genetic algorithms;adaptive operator selection;adaptive parameter control;adaptive pursuit;job-shop scheduling
Código DOI: 10.21528/CBIC2017-30
Artigo em pdf: cbic-paper-30.pdf
Arquivo BibTeX: cbic-paper-30.bib