Hybrid Heuristics Scheduling Algorithm for Hot Strip Mill

This paper introduces the scheduling algorithm of a computer simulation and scheduling system for hot strip mill. The mathematical model of hot strip mill scheduling problem (HSMSP) is formulated. A hybrid heuristics method is proposed, based on parallel strategy, which integrates domain expertise, genetic algorithm (GA) and tabu search (TS). Domain expertise is used to classify and sort production orders to generate initial solutions in parallel. GA is used to improve the initial solutions. And TS attempt to accelerate the convergence of the solutions by adopting swaps of ´single slab´ and ´slab section´. ´Single slab´ and ´slab section´, which we firstly introduced, are new adjacent solutions (neighborhood). Simulation results have shown that the proposed method is feasible and efficient.