Collaborative Composition of Production Services in Multi-agent Systems Based on Auctions

Abstract

Although computer aided design (CAD) and computer-aided manufacturing (CAM) tools have been widely adopted in industry, it still takes a lot of time for the partners to coordinate to create a good solution. To respond to business opportunities, it calls for the development of a methodology to support and automate composition of production processes dynamically. The goal of this paper is to propose a methodology to link process specification models, negotiation mechanism and optimization methods to achieve the desired cycle time and generate processes dynamically. We exploit recent advancements in artificial intelligence and optimization theories to develop a solution methodology for dynamic composition of production processes in multi-agent systems (MAS). To develop such a design methodology relies on an appropriate process specification model to describe the tasks and a mechanism to allocate resources to production processes. Petri nets have been widely applied in modeling of workflows. Combinatorial reverse auctions provide an effective mechanism to select the resources to perform the required operations in workflows. Therefore, we combine Petri net models with combinatorial reverse auctions to dynamically plan the production processes based on MAS and construct a model to control the operations at the shop floor. Our design methodology automates the dynamic composition of production processes. An application scenario has also been provided to verify our solution methodology. We also conduct experiments to illustrate the computational efficiency and scalability of our proposed method.