基于稳定匹配的网约车合乘优化

Ride Sharing Optimization Based on Stable Matching

乘客和司机的选择偏好是影响网约车合乘匹配率的关键影响因素之一。基于乘客对合乘乘客和司机的选择偏好以及司机对乘客的选择偏好,构建嵌入合乘乘客相互影响效应的多对一稳定匹配模型。基于DA(deferred acceptance)算法的思想,设计了考虑同方参与人内部偏好的多对一稳定匹配算法,并通过算例对模型和算法进行验证和分析。结果表明:松弛时间窗下的乘客匹配率平均提高18%,乘客合乘比例平均提升21%。同时,在高需求下,司机匹配率增加,乘客合乘的比例有了大幅提升,司机的收入也有所提高。

The choice preferences of passengers and drivers are one of the key influencing factors on the matching rate of ride-sharing services.Based on the selection preferences of passengers towards passengers and drivers,as well as the selection preferences of drivers towards passengers,a many-to-one stable matching model that embedded the mutual influence effect of passenger sharing was constructed.Based on the idea of DA(deferred acceptance)algorithm,a many-to-one stable matching algorithm that considered the internal preferences over the same party participants was deigned.The proposed model and algorithm were verified and analyzed through numerical examples.The results show that the matching rate of passengers with the relaxed time window increases by 18%,and the proportion of passengers sharing a ride also increases by 21%,averagely.Meanwhile,with high demand,the matching rate of drivers has increased,the proportion of passengers sharing a ride has significantly increased,and the income of drivers has also increased correspondingly.