Optimal path planning for material and products transfer in steel works using ACO

In steel works, steel production from raw material to final product includes several processes. In the processes, pit iron from blast furnace is transported from furnace to storing yards going through factories. For the transportation, because of weight and high temperature of pit iron, torpedo cars towed by a locomotive on a railway in the compounds are used. For effective production, optimal scheduling of transportation is necessary. The transportation includes many constraints; there exist several factories where they should step by, an order among the factories to step by, restricted transportation time, and minimum working time in factories. Hence, to plan the schedule requires complicated calculation and computation time. Also, when an accident occurs and some paths in the plan are closed, they need calculate a new detour in short time period. In this paper, a method to plan schedule for railway train in steel works is proposed. First, a node-arc model of railway systems in steel works is given. Then derived is an optimal path for materials transportation problem. The path goes through specified places in the works in a given order in the shortest steps. In this paper, an algorithm based on Ant Colony Optimization(ACO) is proposed to solve the problem and the ACO method is modified to apply the scheduling problems. The proposed method is applied to a steel works model, and the route generation simulation is given. Finally, the route derived by the proposed method is compared with the route derived by Q-learning method.