An improved double-population genetic algorithm for vehicle routing problem with soft time windows

Abstract

This study primarily focuses on solving the vehicle routing problem with soft time windows (VRPSTW) by applying an improved double-population genetic algorithm (DPGA). The traditional single-population genetic algorithm (SPGA) in solving vehicle routing problem usually traps in local optimum or consumes considerable time. In this paper two different initialization methods — random initialization method and construction initialization method are introduced to frame the improved double-population genetic algorithm. The computation experiment is provided to compare the improved double-population genetic algorithm with the SPGA, and the final outcome effectively proves the superiority of the novel algorithm.