Study On Congestion Facility Location Problem

Because of service capacity restriction and uncertainty of demand process and service time, many facilities are often congested. The study of facility location problems with congestion is a hot point recently. Considering how to make location decision, congested facility location problems are also researched on how to ensure facility service quality,improve customer satisfaction and strengthen facility competition. These decisions interact on each other. From the respective of integration and systematization, this dissertation optimizes the whole system. This dissertation studies congestion facility location problems under different situations based on queuing and location theory. Models are built on different background and many heuristic algorithms are proposed, such as greedy algorithm, lagrangian relaxation algorithm and genetic algorithm. These algorithms are proved to be feasible and efficient.Firstly, this dissertation gives the motivation and significance of congestion facility location research. Based on the current study the disadvantages are concluded, and innovations of this dissertation are presented. Then the main points of congestion facility location problems are given, and the researches to this day are summarized.Secondly, congestion facility locations base on flow interception location problem are studied. Research objects are customer flows traveling on the given paths on the network. The optimization problems under three different backgrounds are studied, which are maximization of expectation interception demands that are lost partially with facilities congestion, two objectives optimization problem with service quantity and quality both satisfied, and competitive peoblem considering customer waiting time. Greedy algorithm and multi-objective evolutionary algorithm are developed based on the characteristic of each model.Thirdly, congestion connection facility location problems with capacity decision are investigated. The distribution function of demand process on each O-D pair is known. Decision variables are location and capacity, and objective function is to minimize the total cost of facility location, traveling, congestion and capacity. Two situations of continuous and discrete capacity variables are respectively studied, and lagrangian heuristics are shown, comparing the difference under these two situations.Fourthly, location problems with time and market impacts are studied. These problems are studied on the base of maximizing covering problem. The location problem with multi-type potential customers is first discussed. With the decisions of target customer selection and facility location, the problem selects customer types according to unit customer profit and average demand rate of each type, and the objective is profit maximizing. Different heuristics are shown, such as neighborhood search, simulation annealing algorithm and genetic algorithm, and different results are compared. Then, facility location problem with promise of response time are studied. Response time includes production time, order waiting time and traveling time. If response time is no more than promise time, customers are satisfied, otherwise, customers are dissatisfied. The objective is to maximize expectation of satisfied customers with budget limitation, and the results of descend algorithm and genetic algorithm are compared.Finally, summarize full dissertation, and present a prospect of future studies.