A sweep-heuristic algorithm combined with partition of time and district for large-scale vehicle routing and scheduling problem of pickup and delivery of customers to airport

Abstract

This paper proposes a sweep-heuristic algorithm combined with partition of time and district to solve large-scale single-trip vehicle routing and scheduling problem (VRSP) of pickup and delivery of customers to airport. Firstly, customers with the same pickup time and pickup location are aggregated into customer orders, which are the basic units for consideration in this paper. The customer orders are further classified into sets according to their pickup time range; these sets are sequenced in ascending order of their pickup time. Secondly, with the depot being the center of all customer orders, the whole service area are divided into circulars according to the driving distance. Every circular indicates different time when car keeps an even speed, and it cuts the service areas into several zones, from the center to the border. Thirdly, every time, we choose customer orders located in the inside circulars from earlier sets and the ones in the outside circulars from later sets to form a new set to be optimized. Finally, generate a radial from the center, and scan the circle with the sweep method to gain the optimized results for this set. Repeat the aforementioned processes, until all customer orders have been serviced. Furthermore, by comparing the results under the criteria of minimum cost and maximum customer satisfaction degree, one can verify that the customer satisfaction degree obtained from the former objective is close to the one from the later objective. However, the cost of the former is apparently lower than the later.