考虑服务半径约束的带预见性竞争选址问题研究

发布时间：2018-01-01 08:41

本文关键词：考虑服务半径约束的带预见性竞争选址问题研究　出处：《清华大学》2015年硕士论文　论文类型：学位论文

【摘要】：在过去这些年,伴随着持续高速增长,中国物流快递行业市场规模已经跃居世界第一位。随着市场的繁盛,每个物流公司都想扩大自己的市场份额,增加公司利润。为了增加企业的市场占有率,给顾客更好的服务感受,增加企业竞争力,某物流快递企业计划建立自营的快递服务便利店,然而其在进行选址决策时知道其它快递企业会在将来不久也进入目标市场,这就是本文将要研究的带预见性竞争设施选址问题。竞争设施选址问题明确地考虑到竞争对手的存在,他们已经(或将要)进入目标市场,而新建设施将与他们竞争获得市场份额。本文将要研究的带预见性竞争设施选址是两阶段设施选址问题,即领导者决策时会考虑到跟随者将来的行为。第一阶段,领导者进行新设施选址决策,使得其市场份额最大,而且领导者知道跟随者的目标函数以及顾客喜好;第二阶段,跟随者已经知道领导者新建设施选址,然后确定其新建设施选址,使得其市场份额最大,同时跟随者也知道各位顾客的喜好。本文在带预见性竞争设施选址问题研究的基础上,结合便利店选址问题实际情况,将设施服务半径约束纳入选址研究范围。首先,基于经典Huff重力模型,考虑服务半径约束,构造了新的吸引力函数,在其中假设设施对顾客吸引力随着他们之间距离的增加逐渐下降,且设施与顾客超出一定的距离时,其吸引力降为零。接着设计出一个相应的顾客选择行为准则:当顾客位于多个设施的服务半径内时,采用随机性模型,即顾客需求在这多个设施之间按照一定的概率分配;当顾客位于单个设施的服务半径内时,采用确定性模型,即顾客需求全部分配给该设施;当顾客超出任意设施服务半径内时,其需求将不能被服务。然后,本文以领导者市场份额最大化为目标函数,按照提出的顾客选择行为规则进行相应的竞争需求分配,构造出一个双层非线性整数规划模型。接着,设计了一种两阶段混合禁忌搜索算法对问题模型进行计算。最后,对设计算法的性能进行测试,通过对小规模算例的求解,并将其结果与最优解进行比较;之后,利用算法对大规模问题进行求解。结果表明算法可以准确获得小规模问题的最优解,而对于大规模问题算法也可以快速进行求解。
[Abstract]:In the past few years, with the continuous rapid growth, the scale of the Chinese logistics express industry has leapt to the first place in the world. With the prosperity of the market, each logistics company wants to expand its market share. In order to increase the market share of enterprises, give customers a better feeling of service, increase the competitiveness of enterprises, a logistics express enterprise plans to establish a convenience store of express service. However, it knows that other express delivery companies will enter the target market in the near future. This is the prospective competitive facility location problem to be studied in this paper. The competitive facility location problem explicitly considers the existence of competitors and they have (or will) enter the target market. New facilities will compete with them to gain market share. This paper will study the location of facilities with foresight competition is a two-stage facility location problem. In the first stage, the leader makes the new facility location decision, which makes the leader have the largest market share, and the leader knows the objective function of the follower and the customer preference. In the second stage, the follower already knows the leader new facility location, then determines its newly built facility location, makes its market share biggest. At the same time, the follower also knows the preferences of customers. Based on the study of location problem with predictive competitive facilities, this paper combines the actual situation of convenience store location problem. Firstly, based on the classical Huff gravity model, a new attraction function is constructed by considering the service radius constraints. It is assumed that the attractiveness of the facility to the customer decreases gradually with the increase of the distance between them, and the facility exceeds a certain distance from the customer. The attractiveness is reduced to zero. Then a corresponding customer selection code of conduct is designed: when the customer is within the service radius of multiple facilities, the random model is adopted. That is, customer needs are allocated according to a certain probability between these facilities; When the customer is within the service radius of a single facility, the deterministic model is adopted, that is, the customer needs are assigned to the facility. When the customer exceeds the radius of any facility service, the requirement can not be served. Then, this paper takes the market share maximization of the leader as the objective function. According to the proposed rules of customer selection behavior, a bilevel nonlinear integer programming model is constructed. A two-stage hybrid Tabu search algorithm is designed to calculate the problem model. Finally, the performance of the design algorithm is tested, and the results are compared with the optimal solution by solving small scale examples. The results show that the algorithm can obtain the optimal solution of the small scale problem accurately, and the algorithm can also solve the large scale problem quickly.
【学位授予单位】：清华大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F259.2

【引证文献】