Agent based material transfer scheduling in steel works

In steel works, locomotives are used to transport products between factories. To attain the stable production, it is necessary to make an appropriate route plan for transferring products between plants effectively and smoothly. However, in a large-scale transportation system of steel works, a lot of products with variety of specification and weight are are treated. As the result, calculation time to make route plan increases, and it is difficult to correspond to frequent transfer request. In this paper, product transfer route planning method in steel works using autonomous decentralized agent is proposed. In the proposal method, each locomotive is considered as an agent, and each locomotive makes the route plan autonomously. In route planning, the problem is to determine transfer route to minimize transportation time with avoiding the collision of locomotives. The entire route plan is cooperatively optimized by each agent repeating an individual optimization of route plan and information exchange.