Applying Distributed Constraint Satisfaction Techniques to Multi-Agent Coordination for Aggregate Planning of Supply Chains

• Main Authors: Yi-Rui PENG, 彭奕瑞
• Other Authors: Yeh-Chun Juan
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/62191238805607281700

碩士 === 華梵大學 === 工業工程與經營資訊學系碩士班 === 97 === Aggregate planning intends to determine the total production level over the planning horizon by resolving the constraints between the product demand and various manufacturing resources for a company and to provide an overall blueprint for ensuing production operations. To enhance the effectiveness and feasibility of aggregate planning, various advanced planning and scheduling (APS) algorithms and systems have been proposed and implemented in many companies. Under the consideration of global manufacturing, aggregate planning has been extended from a single company to the overall supply chain for ensuring the consistency among the aggregate plans from distributed supply chain members. To this end, supply chain members should consider their intra- and inter-aggregate planning constraints and resolve these constraints via supply chain coordination. In this research, an agent-based coordination architecture was proposed for supply chain members to reach a globally accepted aggregate plan. Considering the distributed decision, each agent has an APS module to make its own aggregate planning decision. In addition, a distributed constraint satisfaction algorithm (DCSA) was developed to guide agents through the coordination of constraint conflicts among supply chain members. Finally, a prototype multi-agent system (MAS) is implemented on the Java Agent Development Framework (JADE) platform for the case study of aggregate planning in a TFT-LCD supply chain.