Diffferential evolution algorithm for the multicommodity distribution network design problem | |
| Author | Chayakarn Bamrungbutr |
| Call Number | AIT Thesis no.ISE-11-01 |
| Subject(s) | Business logistics Physical distribution of goods |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Industrial and Manufacturing Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Thesis ; no. ISE-11-01 |
| Abstract | This research presents Differential Evaluation (DE) algorithm to solve multicommodity distribution network design problem (MDNP). This is a strategic planning model where the key decisions include the selection of locations of suppliers, plants and distribution centers (DCs), and to make allocation decisions about the amounts of products from suppliers to plants, plant to DCs and DCs to customers in order to minimize the total cost. A mathematical model is formulated for the 3-stage MDNP (suppliers, plants, DCs) problem. A solution method is proposed based on the differential evolution algorithm by extending the decoding scheme provided by Sae-Huere (2009) to handle the 3-stage MDNP. A new allocation scheme is also proposed based on the highest- demand-to-nearest-DC heuristic. The first experiment is to compare the proposed DE algorithm with existing PSO algorithm using the benchmark data sets for 2-stage MDNP. Based on the experiments, DE algorithm can provide better solutions than PSO algorithm for small and medium size problems. For the large problem instances, the DE algorithm needs more iteration to yield better solutions than PSO. For the 3-stage MDNP, the existing benchmark datasets are expanded to include information on suppliers. In addition, a new customer allocation heuristic is proposed to consider the distance from DC to customer. The heuristic is based on highest-demand-to-nearest-DC decision. Thus, the comparison between these two methods is done for the 3-stage cases. In the 3-stage MDNP, the 2 new customer allocation methods are able to provide better solutions than the customer priority method. Moreover, the allowance of multiple DCs per customer leads to better overall performance. |
| Year | 2011 |
| Corresponding Series Added Entry | Asian Institute of Technology. Thesis ; no. ISE-11-01 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Industrial Systems Engineering (ISE) |
| Chairperson(s) | Voratas Kachitvichyanukul |
| Examination Committee(s) | Huynh Trung Luong;Pisut Koomsap |
| Scholarship Donor(s) | RTG Fellowship |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2011 |