基于物理规划的铝热连轧多目标轧制规程优化

针对铝热连轧轧制规程多目标优化中综合目标函数权值难以确定的问题, 运用基于物理规划的多目标优化方法, 并以萤火虫智能算法为基础, 选取等功率裕量、轧制总能耗、末机架板型良好及各机架打滑因子为目标函数, 进行优化计算。物理规划使设计者以一种更加灵活的方式完成目标函数的权值分配, 然后通过萤火虫智能算法对综合目标函数进行优化计算。结果表明 不同的设计偏好能有效反映设计者的意愿, 从而产生偏向不同重点的轧制规程。

Optimization of Multiobjective Rolling Schedules for Hot Rolling Aluminum Based on Physical Programming

In view of difficulty in determining weights of objective functions in multiobjective optimization for aluminum hot rolling schedule, multiobjective optimization by physical programming method based on the glowworm swarm intelligence algorithm was adopted. Calculation optimization was conducted for the selected objective functions including equivalent rate of power, rolling energy consumption, good shape of end stand and the slip factor. Physical programming was created to ensure weight assignment accomplished in a more flexible framework. Then, comprehensive objective function was optimized by glowworm swarm intelligence algorithm and an optimal rolling schedule was proposed. The results show that, different design preferences can effectively reflect the wishes of designers, therefore different rolling schedule reflecting different priorities can be generated.