基于改进果蝇算法与最小二乘支持向量机的轧制力预测算法研究

摘要

铝合金板材精轧过程中，轧制力是影响板材质量的重要因素。为了满足轧制现场的轧制力预报精度要求，采用改进果蝇算法（ FOA）与最小二乘支持向量机（ LSSVM）相结合进行轧制力预测。改进了果蝇算法的味道浓度判定函数和步长设定方法，采用了分组并行搜索的策略，进而提出一种基于改进FOA-LSSVM的轧制力智能预报方法。将该方法用于铝热连轧现场数据的仿真实验，结果表明样本预测误差在10豫以内，其中84豫的样本误差在5豫以内，精度优于传统模型。%In the process of aluminum alloy plate finishing,rolling force is an important factor affecting the quality of plate. In order to meet the scene of the rolling forecast accuracy,improved fruit flies optimization algorithm( FOA)is combined with least square support vector machine( LSSVM)for rolling force prediction. The function of smell and the method of step length setting are improved. The grouping parallel search strategy is used. A new method based on the improved FOA-LSSVM is proposed for rolling force prediction. The method is used in simulation experiment of field data of aluminum strip,the results show that the prediction error data is in the range of ten percent,and eighty-four percent of these error data is in the range of five percent,the result shows that this method is better than the traditional model.