基于改进蜂群算法的冷连轧规程优化设计

摘要

With certain constraint conditions of facilities on engineering site taken into consideration, a multiobjective optimization model for rolling schedule was established with equalizing rolling load and overcoming slippage as objective functions. The artificial bee colony( ABC) algorithm was modified to improve its performance. Firstly, an initialization strategy based on the opposition⁃based learning was applied to diversify homogeneously the individuals in the search space. Then, several selection strategies were applied through simulation to improve the optimizing accuracy and accelerate the convergence. Finally, schedule optimization strategy for a five⁃stand tandem rolling mill was designed based on the modified algorithm. The results demonstrate that, the modified algorithm can not only avoid effectively the premature convergence, but also improve the overall⁃optimization ability and the convergence speed.%综合考虑现场和设备所受的约束条件，以等负荷和克服打滑为目标函数，建立了轧制规程多目标优化模型。为了提高算法性能，对人工蜂群算法进行了改进。首先，应用反向学习的策略初始化种群，使得个体尽可能均匀分布在搜索空间。其次，人工蜂群算法采用不同的选择机制，提高收敛速度和寻优精度。最后，用改进的算法对某五机架冷连轧机进行规程优化设计。结果表明，改进的人工蜂群算法能有效避免早熟收敛，全局优化能力和收敛速率都有显著提高。