基于改进遗传算法的区域物流网络配送需求预测分析

Forecasting of Distribution Demand of Regional Logistics Networks Based on Improved Genetic Algorithm

针对常规遗传算法的不足,提出在建立区域物流配送需求预测问题数学模型的基础上,构造改进后的遗传算法。应用该数学模型,采用自然数直接编码,对不同种群采用不同选择算子、交叉算子、变异算子,保证了算法的全局收敛性以及收敛速度,同时保证了个体种群的多样性。通过仿真实验对改进的算法与常规的遗传算法进行对比,实验结果表明,改进的遗传算法有较好的全局预测分析能力,提高了算法的计算效率,且算法的全局收敛速度更快,能有效地解决区域物流网络配送需求预测问题。

In this paper, in view of the inadequacy of conventional genetic algorithm, we proposed to, on the basis of the established mathematic model for regional logistics and distribution demand forecasting, add in the improved genetic algorithm. By applying the model, we used natural digits to encode directly and selected different selection, crossover and mutation operators for different populations to ensure global convergency and converging speed of the algorithm as well, at the same time, as the diversity of the populations. Through a simulation experiment in which we compared the modified algorithm with the conventional one, we found that the former was superior in all aspects with regard to the latter.