A multiagent systems approach for managing supply-chain problems: a learning perspective

The main purpose of this paper is to assess if multiagent systems (MAS) technology can be used to model, study and manage supply chains (SC). A second objective is to describe the application of the Q-learning algorithm and present some results obtained afterwards. To validate this hypothesis it was modelled and implemented a MAS with the various functionalities, being learning one of them. Two examples, one SC of the chemistry industry and a peer to peer, are presented to evaluate the usefulness of the MAS technology and the impact of learning on these systems.