Multi-objective Particle Swarm Optimization with Kriging forSupply Chain Inventory Problems
碩士 === 國立高雄第一科技大學 === 運籌管理研究所 === 99 === This paper intends to improve the research by Hou (2010) research, which combines an MOPSO algorithm with a simulation model to solve a supply chain inventory management problem. The algorithm implements a technique called Kriging to forecast the performance...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2011
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/08455701546876184847 |
| Summary: | 碩士 === 國立高雄第一科技大學 === 運籌管理研究所 === 99 === This paper intends to improve the research by Hou (2010) research, which combines an MOPSO algorithm with a simulation model to solve a supply chain inventory management problem. The algorithm implements a technique called Kriging to forecast the performance of each particle at the new position so that no simulation is required for solutions appear to be uncompetitive. In order to improve the convergence and/or diversity of the Pareto set, this study applies Kriging to explore the neighborhood of each non-dominated solution. Then the new algorithm revises and obtains an expanded set of non-dominated solutions in which a leader (Gbest) is selected for each particle which may guide particles moving toward better solutions. Numerical testing results suggest that the new MOPSO algorithm can achieve significant improvement.
The supply chain management problem is about the operations of a supplier hub operated by a third-party logistics provider. Ordering and inventory decisions will affect service level and operating cost of the hub. Under the VMI proposal, the 3PL can access inventory information of all parties and is responsible for making all decisions. Our algorithm constructs a Pareto-optimal front between inventory costs and fill rate to evaluate VMI benefits. The experimental results suggest that the VMI model can achieve the same fill rate with less inventory cost. In addition, VMI advantage is more evident when the number of manufacturers involved increases.
|
|---|