Multi-Behavior Agent Model for Supply Chain Management (Track 1, Session 12)
Author(s) :
Pascal Forget (Laval University, Canada)
Sophie D'Amours (Laval University, Canada)
Jean-Marc Frayret (Laval University, Canada)
Abstract : Recent economic and international threats to occidental industries have encouraged companies to rethink their planning systems in a way to quickly react to and correct deviance from established plans, respond to demand, reduce inventory and exchange information promptly throughout the supply chain, in other words become more agile. Due to consolidation, the development of integrated supply chains and the use of inter-organizational information systems have increased business interdependencies and the need for collaboration. Thus, agility and the ability to deal quickly with disturbances in supply chains are critical to maintain overall performance. In order to develop tools to increase the agility of the supply chain and to promote the collaborative management of such disturbances, agent-based technology takes advantage of the ability of agents to make autonomous decisions in a distributed network. This paper proposes a multi-behavior agent model geared with tools designed to improve agility in supply chains, which presents different behaviors to plan manufacturing activities and deal with disturbances in a distributed collaborative context. This agent model uses different behaviors following different situations, using standardization, utility evaluation, anticipation mechanisms and negotiation protocols. Depending on the environment and the type of disturbance, the agent can select the preferred behavior. A behavior scenario involving a specific disturbance is presented and an implementation strategy is proposed, with the objectives to prove feasibility and increase supply chain performance. The implementation of this solution is realized through the FOR@C experimental agent-based planning platform, dedicated to the supply chain planning for the forest products industry. The North American lumber industry represents a perfect context for this technology, mainly due to the large amount of stochastic disturbances in many aspects of the supply chain.

Menu