基于改进多目标粒子群算法的冷连轧规程优化设计

合理的轧制规程能够提高轧机的产量和产品的质量，带来显著的经济效益。采用多目标粒子群算法，选择等相对负荷和预防打滑为目标进行冷连轧规程优化。针对算法存在的收敛性和分布性难以均衡的问题，引入一种基于平行坐标系的密度和收敛潜能计算方法;同时，为克服算法易于陷入局部最优的缺陷，提出一种带个体扰动的全局最优领导粒子选择策略。仿真结果表明，该方法能快速跳出局部极值，获得具有更好收敛性和分布性的近似Pareto前沿。最后应用该方法对某五机架冷连轧机进行了轧制规程优化。 

Optimization of Tandem Cold Rolling Schedule Based on Improved Multi-objective Particle Swarm Optimization Algorithm

Reasonable rolling schedule could improve the quality of the mill's production and products thus bring significant economic benefits. The multi-objective particle swarm algorithm was adopted to optimize the objective functions of equal relative load and the slip rate. According to the problems that convergence and distribution were difficult to balance, based on parallel cell coordinate system a novel method was proposed to assess the evolutionary environment including density and potential of convergence. Meanwhile, to enhance the ability of escaping from local optima, a disturbance vector strategy was proposed for selecting global best. Simulation results show the improved strategy can be able to prevent falling into local extremum point, and improve the convergence and distribution of approximate Pareto front. The method was applied to a five-stand tandem cold mill rolling schedule optimization.