浸出过程浸出率预测与在线优化

摘要

本文建立了浸出过程浸出率混合预测模型；基于该预测模型及生产的实际需要,将一个动态浸出问题转化为带约束的优化问题；针对在线优化计算时间长,精度要求高的难题,提出了变邻域搜索PSO算法及基于该算法的动态生产过程在线优化策略.通过实际应用证明混合预测模型预测精度高；在线优化算法自适应性强、运算量小、精度高,具有较高的实际应用价值.%A hybrid predictive model for the leaching process is proposed. The dynamic leaching problem is formulated as a constrained optimization problem, based on the predictive model and the practical demands of manufacture. Aiming at the characteristics of long computation time and high demand of predictive precision in online optimization, a PSO optimization algorithm with variable neighborhood search is introduced, and an online optimization strategy of dynamic manufacture process is proposed based on the algorithm. The practical operation shows that the predictive precision of the hybrid predictive model is higher and the online optimization algorithm is self-adaptive which can achieve obvious practical value with less calculation quantity and higher precision.