Solving the simple plant location problem by genetic algorithm

From MaRDI portal





The authors have tried to explore limits of genetic algorithm employed for solving hard integer programming problem. The suggested algorithm belongs to family of metaheuristics working with a set of feasible solutions considered as a population of individuals. The algorithm simulates development of a population in the nature by resembling of selection, crossover and mutation. Algorithms of this family can avoid a local minimum trap but it is commonly known that their convergence is very slow. To study properties of the algorithm, the well known simple plant location problem was chosen. This problem constituted object of researchers interest in the past twenty years and this way the previous studies provide not only considerable set of benchmarks but even methods producing exact solution of the problems, making use of these preliminaries the authors investigated behaviour of their algorithm. They compared it with Erlenkotters exact approach and showed that their metaheuristic employed for very large instances of the problem yields an optimal solution with the same frequency as the exact approach, but in a fraction of the associated time.




Cited in
(32)








This page was built for publication: Solving the simple plant location problem by genetic algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2765605)