基于量子位实数编码的优化算法及轧制规程多目标优化

An optimization quantum-bit real-coded algorithm for optimizing multi-objective rolling schedule

针对热连轧轧制规程优化问题,以等功率裕量和轧制能耗为优化目标函数建立热连轧轧制规程多目标优化模型,提出基于量子位实数编码的热连轧轧制规程多目标优化算法。该算法将免疫遗传算法框架与量子计算思想相结合,采用量子位实数编码,利用量子态干涉进行遗传算子的交叉和变异,同时保证非支配解按拥挤距离选择优势免疫抗体种群,得到Pareto全局最优解集。以某轧钢厂热连轧精轧机组为例,验证本文所提及算法的有效性。实例分析表明,所提及的算法在寻优能力和收敛速度上均优于传统的NSGA-Ⅱ算法,能够获得更好的Pareto解集,有效地解决热连轧轧制规程多目标优化问题,改善了轧制能耗。

To optimize rolling schedule in hot continuous rolling, a multi-objective optimization model with optimized objective function based on equal power allowance and rolling energy consumption is constructed, and a multi-objective optimization algorithm based on quantum-bits real-coded is proposed for rolling schedule in a hot continuous rolling. The algorithm combines immune genetic algorithm framework with quantum computing idea, employs quantum-bits real coding, interferes crossover and mutation of genetic operator using the quantum state, and ensures the non-dominated solutions selecting advantage immune antibody population based on crowded distance so as to obtain the optimal Pareto solutions. Taking the finishing mill group in hot continuous rolling as an example, the effectiveness of the proposed algorithm is verified. The example analysis indicates that the optimization ability and convergence speed of the proposed algorithm are better than the traditional NSGA-II algorithm, which can obtain better Pareto solutions. The algorithm effectively solves the multi-objective optimization problem for the rolling schedule and improves the rolling energy consumption.