ABSTRACT
If the product supply shortage occurs during the sales period, customers will turn to other companies, so the enterprise lose the sales opportunity. If the enterprise can predict product demands, and manages the product sales by using the supply chain management prediction system to improve the bottleneck of the inventory, the hot sales product can have more critical time utilization, and the inventory status can be reflected quickly by Internet of Things (IoT). To overcome the problem that the replenishment model cannot show the actual quantity of products on the store shelves, in the project, we propose an intelligent agent-based prediction system, which serves as a framework to construct an integrated prediction system through the use of radio frequency identification (RFID) technology to design the intelligent product prediction shelf to extract product messages, and the service oriented architecture to develop prediction information to recommend products to the customer. The result of the project proposes an agent-based cloud computing service platform in IoT and intelligent agents with SOA as backend cloud services. To build a prototype prediction system with performance analysis, it can be proved that the prediction system architecture for intelligent agent-based prediction system could improve operation performance and effectively enhance customer service quality for hot sale products.