Jozef Kratica
Vladimir Filipović,
Vesna Šešum
Prof. Dušan Tošić
The uncapacitated warehouse location problem is considered. Since it belongs to the class of NP complete problems, we use the genetic algorithms in the solving of this problem. Genetic algorithms are rooted in the mechanisms of the evolution and natural selection. They are relatively general and practicale way for the finding a suboptimal solution (heuristic) in the problems of optimization.
According to the uncapacitated ware-house location problem, we should find provision plan with minimal total cost. The storage cost for every warehouse and the cost of shipment from every warehouse to an arbitrary customer are known.
We use simple genetic algorithm for the solving of uncapacitated warehouse location problem.. The item-code is represented by the binary array of indicators denoting the inclusion of warehouse into provision plan.
This approach seems to be a good compromise between the quality of solution and execution time. The improvements are possible by introducing of other selection and crossover operators.
Keywords: Uncapacitated warehouse problem, NP complete problems, Genetic algorithms
Back to Jozef's Papers Home Page | Back to Jozef's Home Page | Full paper - PDF (107 KB)