基于混合遗传算法的模糊需求车辆路径问题

针对模糊需求的车辆路径问题,以模糊可信性理论为基础,设计了混合遗传算法。引入扫描算法进行种群的初始化,结合配送分队数和剩余载重因素提出了混合交叉算子。借鉴差分进化和扫描思想构建了基于整数序规范的差分扫描变异运算。以Solomon标准集为算例,通过实验确定了主观参数的最优取值为0.4。改进前后的阶段效果比较表明,混合交叉算子和差分扫描变异算子提高了算法的性能,采用可信性理论研究模糊需求问题有一定的优越性。

Vehicle routing problem with fuzzy demands based on hybrid genetic algorithm

A hybrid genetic algorithm is designed based on fuzzy credibility theory according to the vehicle routing problem with fuzzy demands. Firstly sweeping algorithm was introduced to the population initialization. Secondly a hybrid crossover operator was proposed combined with the number of residual demands and distribution units. Then difference sweeping mutation operation was constructed based on integer sequence criterion by differential evolution and sweeping thinking. As Solomon standard sets an example, the subjective parameter is determined the optimal value of 0.4 by the experiment. The stage effects before and after the improvement show that the hybrid crossover operator and the differential sweeping mutation operator have improved the performance of the algorithm, and credibility theory demand has certain superiority to the fuzzy demand problem.