关键词: 轴-辐式网络, 竞争, 服务约束, 连续化方法, 增广Lagrange乘子罚函数法

Abstract: The current trend of global trade enhance the importance of shipping service, since it is in charge of transporting up to 90% of the trade volume. Recently, a number of shipping alliances have emerged to dominate the shipping market, and most of the smaller shipping companies are suffering from an ongoing loss of their profits. It is therefore essential for the smaller shipping companies-that are not operating in shipping alliances to seek for ways of competing with the shipping alliances in order to achieve some higher profits or at least to recover the lost market shares. An integer nonlinear programming model is propesed for the hub-and-spoke shipping network optimization with service constraints in a competitive environment to address the considered problem. An existing shipping alliance, called the leader, utilizes a transportation network with a multi-allocation hub-and-spoke topology. A new shipping company, the follower, wants to offer its shipping service in the same shipping market, using its own multi-allocation hub-and-spoke shipping network and setting service quality, service time and service cost so as to maximize its profits. The question to be answered is: Can the follower obtain profits under these conditions, even with same service quality, service time and service cost ofthe leader? In order to answer this question, our procedure finds how many hub ports to locate, where should they be located, what is the best route network. The contributions of this paper are as follows. In the first place, continuous hub location model (the domain of hub ports is a plane not a series of particular ports) is formulated. Secondly, the numbers of routes existing in the origin-destination ports are extended. Third, an attraction function which is a proportional model not a discrete choice model is provided to simulate the consignors' choice behavior. Finally, the integer nonlinear problem is solved using an augmented Lagrange function method based on NCP function and coagulation function. Consequently, the conclusions are achieved by example simulation that, (1) the follower will obtain certain profits by opening moderate number of hub ports in the case of service cost is considered by consignors (θ>0), even if there is no economies of scale (α=1.0); (2) the follower's benefits will be the most significant if there are high economies of scale (α=0.2), but its profits in the case of the leader has different amount of hub ports (P_A) located will not unified converges to a certain value if there is no economies of scale (α=1.0) by the fact of a proportional model is applied; (3) the follower can obtain much profits by opening more hub ports if the leader has one hub ports (P_A=1), but its capability of obtaining a higher profits will be reducing if the leader has operated more than two hub ports (P_A>2) for the 12-node versionof the shipping network.

Key words: hub-and-spoke network, competition, service constraints, continuous approaches, augmented Lagrange function method