Distribution Maintenance Time scheduling Using a Genetic Algorithm
Xianchao Huang, Lizi Zhang, Jun Shu, Jingwei Zhang · 2008
Distribution maintenance time scheduling is a multi-objects and multi-constrains optimizing problem. To solve this complicated problem, this paper presents an improved genetic algorithm that developed "direct compare algorithm" in constructing fitness function and mutation operator of simple genetic algorithm. By introducing "infeasible area", the improved algorithm allows individuals among the infeasible area compete with their target values. So a part of infeasible individuals with good gene can be reserved and the performance of the optimizing algorithm has been improved. At the same time, the improved algorithm designs a new mutation operator which mutates by tabu searching when the population is multiplex enough and changes its mutates way to uniform automatically when the average population multiplex lower down to a fix degree. The tabu searching mutation operator can speed up constringency of the algorithm and the uniform mutation operator can bring new gene to avoid "premature" of genetic algorithm. The proposed method is applied to a practical system, and numerical results verify the correctness and validity of it.