基于改进差分进化算法的出租车合乘问题研究

在现有出租车合乘问题研究基础上，考虑乘客模糊时间窗、合乘意愿等因素，建立 更符合实际的多对多合乘模型，并使用改进的差分进化算法进行求解.该算法设计了一种分段 实数编码方案、基于个体排序的缩放因子F 与交叉概率CR及混合轮盘赌的半贪婪选择策略. 仿真结果表明：所提算法的求解效果优于传统差分进化算法与遗传算法，是解决该类问题的 有效方法；与非合乘模式相比，所建模型成本减少、服务乘客数增加，模型合理有效；同时分析 了模糊时间窗、合乘意愿2个因素对模型求解结果的影响.

Shared Taxi Problem Based on the Improved Differential Evolution Algorithm

A more realistic many-to-many dynamic shared taxi model is established, with considering the factors such as fuzzy time window and willingness to carpool, and the improved differential evolution algorithm is supposed. A segmented real number coding scheme is designed in this algorithm, as well as the scaling factor F and the crossover probability CR based on the individual ranking and a semi- greedy selection strategy mixed roulette. The simulation results indicate that the proposed algorithm could find higher quality solutions than the traditional differential evolution algorithm and the genetic algorithm; compared with the non- carpool mode, the total cost of the developed model is reduced, and the number of served passengers is increased; besides, the influence of the fuzzy time window and willingness to carpool on the solution of the model is also analyzed.