基于并行遗传算法的车辆路径问题

Solution to Vehicle Routing Problem Based on Parellel Genetic Algorithm

提出两种改进策略来提高遗传算法的性能,首先通过粗粒度并行机制以避免遗传算法在进化过程中易产生过早收敛现象,同时提出了一个主从式迁移策略来提高"优质"个体在交换过程的生存能力,有效的提高优化的速度和解的精度。最后,通过若干著名的车辆路径问题对该算法进行了验证,结果表明提出的并行遗传算法可以有效的提高优化速度和求解质量。

The paper proposes two improvement strategies for the genetic algorithm （GA） in logistics research. The first strategy is the coarse-grain parallel mechanism which can enhance the solution quality by exchanging high-quality genus among sub-colonies, thus avoiding premature convergence in the iteration process and the second one a primary-subordinate migration strategy which potently improves the speed of optimization and the accuracy of solution. Finally, through typical vehicle routing problems, the algorithm proposed in the paper is validated as effective in improving the speed and quality of GA.