A multi-agent approach for dynamic production and distribution scheduling

Abstract

In an age when enterprises are increasingly dependent on their suppliers and customers, individual companies should broaden their scope in order to include other organisations participating in the same value chain. Consequently, different parties of the chain ought to exchange information to make sound decisions in order to cut down global costs or improve customer service. This paper proposes a multi-agent approach for dynamic production and distribution scheduling in a simple supply chain. The approach is based on the continuous supervision of the active schedules and routes, in order to detect possible exceptions and apply corrective actions in a real-time and coordinated manner with other parties. Finally, the paper focuses on the design phase of the distribution stage, where the problem is mathematically formulated as a dynamic vehicle routing problem with time windows and backhauls (DVRP-TWB), and three combinatorial optimisation meta-heuristics are proposed as solving techniques.