| We consider the express delivery problem with multiple vehicle types in urbanarea, a variant of the vehicle routing problem in which the customers can be classifiedinto two types, one is deterministic with time window, while the other appearprobabilistically with stochastic demand. The vehicles also have two types, which aredifferent in capacity and variable cost. In Chinese express delivery service, we callthem “cavalry” and “infantry”. Each type of vehicles can serve only one type ofcustomers. However, the most important is that cavalry can replenish express toinfantry. Thus infantry which has smaller capacity can continue service without goingback, which will reduce the travel cost. Exchange between vehicles is a new modewhich is very useful in reality, but is rarely studied.The primary objective of this research is to provide a realistic model to minimizethe travel cost when customer location and demands vary from day to day whilemaintaining drivers’ familiarity with the routes and make the same customer to beserved by the same courier at almost the same time whenever he requires service. Wecreated a two-stage vehicle routing model-called periodic stage and daily stage. Wepropose the concept of service cell, obtained through clustering method to balance theobjective of maintaining routes familiarity and serve varying customers every day. Inorder to model the exchange process, we select a service cell as the exchange cell andcreate the exchange strategy. The periodic strategy generates the plan about servicearea and exchange, while the daily schedules optimize the real service routes when allthe customer information is obtained before the day.In order to solve large-scale problem, we solve the model in two stages. Asimulation model was built to test the model. We analyze effect of clustering methodto the service area, the exchange strategy and the optimization result. The experimentshow good result on solving the express delivery problem effectively and efficiently. |
PDF Full Text Request