基于乘客出行偏好的网约车合乘模型研究  

Ridesharing model for online ride-hailing based onpassengers’ travel preferences

作　　者：钱光武 金典 向宣奕 QIAN Guang-Wu;JIN Dian;XIANG Xuan-Yi(Sichuan University-Pittsburgh Institute,Sichuan University,Chengdu 610207,China)

机构地区：[1]四川大学匹兹堡学院,成都610207

出　　处：《四川大学学报(自然科学版)》2025年第2期433-442,共10页Journal of Sichuan University(Natural Science Edition)

基　　金：国家重点研发计划(2024YFC2510704)。

摘　　要：针对网约车合乘中忽视乘客绕行容忍度而导致合乘体验和意愿下降的问题,本文引入模糊绕行窗的概念,并将其与模糊时间窗相结合来量化乘客的个性化需求,构建了一种能够同时满足乘客的上车时间偏好和绕行容忍度的网约车合乘模型.该模型以最小化行驶距离、乘客费用和惩罚函数为优化目标,综合考虑乘客的绕行窗、时间窗以及车辆额定载客量等约束条件.基于乘客公平性原则提出了一种费率分摊方法,对乘客的费用和驾驶员的收益进行约束,实现司乘双赢.结合模型特点,设计了改进的遗传算法对优化模型进行求解.实验结果表明,所得最优合乘方案相较于单乘模式,承载的乘客数量提高了33.3%,乘客的出行费用降低了24.3%,同时满足了乘客的个性化需求.Addressing the issue of neglecting users'tolerance for detours in ridesharing,which directly impacts their ridesharing experience and participation willingness,this paper introduces the concept of fuzzy detour window and integrates it with fuzzy time window to quantify users'subjective travel requirements.The proposed approach establishes a comprehensive ridesharing model that incorporates pickup time preferences and detour tolerance constraints.In this model,a multi-objective function is constructed to minimize travel distance,user costs,and penalty functions,subject to constraints such as detour windows,time windows,and vehicle capacity limits.To ensure mutual benefits for all participants,an equitable cost allocation mechanism is developed,effectively balancing user costs and driver revenue.Finally,an improved genetic algorithm tailored to the ridesharing problem is proposed to solve the optimization model.Experimental results show that,compared to single-occupancy trips,the optimized ridesharing solution achieves a 33.3%increase in passenger capacity and a 24.3%reduction in user travel costs,while satisfying personalized travel needs of users.

关 键 词：合乘 遗传算法 多目标优化 个性化需求 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]