Integrated use of soft computing and clustering for capacitated clustering single-facility location problem with one-time delivery

• Main Authors: Liu, Hsiu Hao, 劉修豪
• Other Authors: Yeh, Wei Chang
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/44605285239031145119

碩士 === 國立清華大學 === 工業工程與工程管理學系 === 103 === How to set up the distribution centers not only a very general and important issue but also a quite profound knowledge for many different industries in real life. This kind of problem called capacitated clustering location problem (CCP), then the previous literature, the target of the location problem is to minimize the sum of distances from each cluster centers to all customers in their cluster and the practical application of this type, for example: logistics centers, police stations, etc., this approach of shipping called “batch delivery”. However, some types of business or the company shipping methods are different from the above. The distribution centers must ship goods to all customers in its cluster or cluster in a cycle time, such as postman, newsman, etc., this approach of shipping is called “one-time delivery”. One-time delivery can be modeled as a Travelling Salesman Problem (TSP). This type of problem is rarely discussed in the former literature. In order to solve this practical problem, a new hybrid clustering algorithm, Simplified Swarm Optimization combining with K-harmonic means (SSOKHM) is proposed in this paper for the CCP with one-time delivery and using greedy algorithm for the TSP to find the shortest loop and obtain the target: minimize shipping costs. An exchange local search is an inspection mechanism which adds after greedy algorithm. Proposed method is applied to several facility location problems from OR library and compared with other hybrid clustering algorithm. Numerical results show that the proposed approach (SSOKHM) performance is better than using other hybrid clustering algorithms in terms of shipping costs. Finally, we embed the exchange local search for each hybrid clustering algorithm. The results are presented that this inspection mechanism can be work in capacity constraint to enhance the performance of the solution and demonstrate the usefulness of this approach in CCP. Keywords: Capacitated Clustering Location Problem, Simplified Swarm Optimization (SSO), K-harmonic Means (KHM), Travelling Salesman Problem