基于自适应进化算法的烧结配料优化

Optimization of Sintering Blending Based on Adaptive Evolutionary Algorithm

针对烧结配料系统的多样性、复杂性和相关性特点,基于广义回归神经网络建立了烧结配料预测模型,提出基于自适应加速机制的多种群进化算法的烧结配料优化算法。该算法在引入自适应加速机制和弹性缩放因子的前提下,充分运用了多种群进化算法的全局搜索能力寻找最优的工艺参数组合,将神经网络和自适应进化算法有机结合,实现了烧结配料的优化,增加了混合料中有用的化学成分,从而提高了产品质量。实际计算结果验证了该优化算法的正确性。

A prediction model of sintering blending was established based on GRNN according to characteristics of diversity, complexity and relativity of sintering blending system, and a new optimal algorithm was proposed based on multi-group evolutionary algorithm with adaptive acceleration mechanism. The algorithm uses global search ability of multi-group evolutionary algorithm to achieve the optimization of process parameters under introduing adaptive acceleration mechanism and elastic scaling factor. The organic combination of GRNN and adaptive evolutionary algorithm is helpful to achieve proportion optimization of sintering blending and increase useful chemical composition in sintering blending, so as to improve product quality. The correctness of the optimization algorithm was verified with practical calculation results.