改进混合粒子群算法优化热轧负荷分配

Optimization of hot rolling load distribution by improved hybrid particle swarm optimization algorithm

热轧带钢精轧过程的负荷分配是制定精轧轧制规程的核心。为制定合理的负荷分配方案并提高生产效率与产品质量,在粒子群算法中引入杂交策略和收缩因子,构成改进后的混合粒子群优化算法。杂交策略的引入可以增强粒子种群的多样性,避免算法在收敛后期因粒子多样性不足而无法跳出局部极值。收缩因子可以有效约束粒子飞行速度,增强算法局部收敛能力。杂交策略与收缩因子两者共同作用可以明显改善算法的收敛性能。函数测试与仿真试验结果表明,改进的算法具备很好的收敛速度和收敛精度,对负荷分配的优化也十分有效。

The load distribution in the finishing process of hot rolled strip is the core of making the finishing rolling schedule.In order to make a reasonable load distribution scheme and improve production efficiency and product quality, hybridization and contraction factor were introduced into particle swarm optimization to form the improved hybrid particle swarm optimization.The introduction of hybrid strategy could enhance the diversity of particle population and avoid that the algorithm couldn’t jump out of local extremum due to the lack of particle diversity at the later stage of convergence.The contraction factor could effectively constrain the particle’s flight speed and enhance the local convergence ability of the algorithm.The convergence performance of the algorithm could be improved obviously by the combination of hybridization and contraction factor.The results of function test and simulation show that the improved algorithm has good convergence speed and accuracy, and it is also very effective for the optimization of load distribution.