首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于Mixed Logit模型的在线订单配送时隙动态定价
引用本文:马文景,陈淮莉.基于Mixed Logit模型的在线订单配送时隙动态定价[J].上海海事大学学报,2018,39(4):51-57.
作者姓名:马文景  陈淮莉
作者单位:上海海事大学物流科学与工程研究院,上海海事大学物流科学与工程研究院
基金项目:国家社会科学基金(15BGL084);上海市科学技术委员会科研计划(14DZ2280200);上海市哲学社会规划课题(2014BGL018)
摘    要:为研究电商环境下配送时隙的动态定价决策优化问题,基于收益管理的思想,考虑到客户对时隙选择偏好不同且具有随机性的特点,在Mixed Logit客户选择概率模型中引入效用函数,建立配送时隙的期望收益模型,并采用强化学习算法寻求时隙动态定价优化策略。通过算例分析验证模型的有效性。结果表明:Mixed Logit模型的拟合效果优于多项式Logit (multi-nomial Logit,MNL)模型的拟合效果,其灵活的参数分布能更好地处理客户偏好的随机性,从而更准确地描述客户的时隙选择行为;用强化学习算法求解模型时,Mixed Logit模型比MNL模型的求解速度更快、结果更优。

关 键 词:配送时隙    Mixed  Logit模型    强化学习    动态定价
收稿时间:2017/11/23 0:00:00
修稿时间:2018/1/24 0:00:00

Dynamic pricing of delivery time slot for online orders based on Mixed Logit model
Institution:Shanghai Maritime University Logistics Research Center
Abstract:In order to study the dynamic pricing decision optimization issue of delivery time slots in e commerce environment, based on the theory of revenue management, the utility function is introduced into the Mixed Logit customer selection probability model, where the different preferences and randomness of time slot selection of customers are considered. The expected return model of delivery time slots is established, and the reinforcement learning algorithm is used to find the dynamic pricing optimization strategy. The validity of the model is verified by an example. The results show that, the fitting effect of the Mixed Logit model is better than that of the multi nomial Logit (MNL) model, and its flexible parameter distribution can better deal with the randomness of customer preference and describe time slot selection behavior of customers more accurately. When the reinforcement learning algorithm is used to solve the model, the Mixed Logit model has faster calculation speed and better result than the MNL model.
Keywords:delivery time slot  Mixed Logit model  reinforcement learning  dynamic pricing
本文献已被 CNKI 等数据库收录!
点击此处可从《上海海事大学学报》浏览原始摘要信息
点击此处可从《上海海事大学学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号