An Adaptive Genetic Algorithm for Solving Practical Scheduling with Maintenance Plan and Different Accessibility Relations in Steel Plants

Since the hot metal pretreatment-steelmaking-continuous casting (PSCC) process has become the most common process in modern steel plants, the traditional steelmaking-continuous casting (SCC) scheduling is being extended and transferred to the more complex PSCC scheduling. In this paper, based on the existing studies for SCC scheduling, the maintenance plan and different accessibility relations between machines are considered in the PSCC scheduling to satisfy the practical production requirement in modern steel plants. A heuristic that contains a machine allocation procedure and a conflict eliminate procedure is developed first to obtain a good feasible schedule of this problem. Subsequently, by using the heuristic as a decoding algorithm, an adaptive genetic algorithm (AGA) featuring a novel adaptive evolutional strategy is presented to obtain a suboptimal solution of the PSCC scheduling problem. Experimental results based on instances generated from a realistic steel plant demonstrate that the AGA is effective for solving PSCC scheduling with maintenance plan and different accessibility relations.