| Summary: | 碩士 === 雲林科技大學 === 工業工程與管理研究所碩士班 === 96 === In recent years, an increasing number of consumers are considering not only the price factor but also the after-care service when buying a product. Therefore, for enhancing the market competitiveness of a product, the producer would anticipate the after-care service mechanism before designing and manufacturing it.
The goal of this study is proposing a multi-objective mathematical model for getting the minimum total cost and the difference of average utilization rate between maintenance facilities, which can provide some references for the industry. In this study, the evacuation probability is added into the model, which is the possibility of the products that can not be fixed completely by general maintenance stations and needs to be sent to professional ones and repaired.
In the allocation of the model, when only consider about the total cost, it is the centralized strategy; when only consider about the the difference of average utilization rate between maintenance facilities, it is the decentralized strategy.
In the solution of the model, an improved immune algorithm is proposed in this study, which introduces the mechanism of Hamming distance, and deletes the dominanted solution and elite policy. Then, the solution of test samples is carried out, and it is compared with the immune and genetic algorithms. According to the results, it is found that the quality of solution in improved immune algorithm is better than that in immune and genetic algorithm. Moreover, for the genetic algorithm, the difference of solutions between two algorithms is reduced when the scale of the problem is smaller; but for the immune algorithm, the difference of solutions between two algorithms is increased when the scale of the problem is larger..
Finally, the analyses of sensitivities for evacuation probabilities, number of initial population and termination generations are executed. The results show that the evacuation probabilities influence the total cost, which means the number of professional maintenance stations becomes less when evacuation probabilities decrease. On the other hand, when the evacuation probabilities increase, the needed number of professional maintenance stations becomes more. Also, the distance between general and professional maintenance stations will decrease because of the reduction of transportation costs.
According to the results of analyses, it is found that the number of initial population and termination generations will influence the quality and speed of solution. With regard to the qualities of solution, the influence is less when the scale of the problem is smaller. On the contrary, the influence is greater when the scale of the problem increases, which illustrates that the qualities of solution is better when the number of initial population and termination generations are greater. No matter what the scale of the problem is, it has a great influence on the speed of solution. And, it spends more time for solving process when the initial population and termination generations are greater.
In terms of the results, the study proposes a good design of reverse logistics network, which gives consideration to both quality of customer service and capacity utilization of maintenance center. It can provide the industry with the reference for practical design of service network, which makes the design of logistics network more faultless. It is hoped to create a win-win situation for both retailers and customers, which is the goal and contribution of this study.
|
|---|
