基于模糊聚类和识别的出租车合乘算法研究

针对目前部分大城市出租车合乘效果差，合乘效率低等现状，本文采用模糊聚 类和模糊识别方法，研究出租车行驶路线模糊聚类，并利用行驶路线、行驶时间和合乘人 数创建隶属函数,实现合乘乘客与出租车的合乘模糊识别.通过随机生成的多组出租车出 行和合乘乘客样本数据，发现在假定的出租车合乘条件下，出租车样本数量决定了合乘 的成功率，但同时也发现，在合乘人数固定的情况下，无限制的增加出租车样本数量会增 加合乘乘客的搭载成功率，平均每辆合乘出租车的收入并不会随着样本数量的增大而增 大，而是趋于稳定值.仿真结果说明，该算法适合于大样本的出租车合乘问题，是一种可以 提高出租车合乘成功率的有效方法.

Algorithm Research of Taxi Carpooling Based on Fuzzy Clustering and Fuzzy Recognition

In some big cities，the effect and efficiency are both poor to the taxi carpooling. The taxi route clustering and carpooling identification of passenger and taxi are studied by fuzzy clustering and fuzzy recognition theory. Through randomly generated many groups of taxi and passengers data, it is pointed that taxi carpooling in particular conditions, taxi numbers will decided the passenger’s carpooling success rate, but it is found that unlimited increase taxi sample number will increase carpooling success rate to passengers in case of a fixed number of passengers, the taxi income will be stabilized, it will not increase with rising of taxi number. The results indicate that the algorithm is suitable for us take the carpooling problem of large numbers taxi and could be effective measurement for taxi carpooling.