改进果蝇算法在净化除钴过程锌粉量优化设定中的应用

本文提出一种基于细菌迁徙的自适应果蝇算法(CAFOABM),用于求解净化除钴过程锌粉添加量优化设定问题.首先从净化除钴的反应机理出发,通过引入除钴率建立锌粉添加量与氧化还原电位值、出口钴离子浓度间带约束的非线性优化设定模型;为了避免果蝇算法在迭代寻优过程中出现停滞现象,CAFOABM引入多种群并行搜索策略和改进搜索算子对搜索状态的转移规则进行改进;并采用区分可行解与不可行解法进行约束处理,保留一定比例的不可行解个体,有效避免了有用解信息的丢失.实际生产数据验证结果表明,CAFOABM算法优化设定的锌粉添加量与人工操作的经验数据相比减少7.83%,出口钴离子浓度满足实际生产要求且趋于平稳.

Application of improved fruit fly optimization algorithm in zinc powder dosage optimization of cobalt removal process

This paper proposes an adaptive fruit fly optimization algorithm based on the bacteria migrate (CAFOABM) to solve the zinc powder dosage optimal setting problem of cobalt removal process. Firstly, removal rate of cobalt is introduced as an intermediate variable to help formulating the relation models between the zinc powder dosage, oxidation reduction potential (ORP) and outlet cobalt ion concentration. This nonlinear optimal setting model is carried out based on the reaction mechanism of cobalt removal process. In order to avoid the stagnation of fruit fly system in iterative optimization process, parallel strategy of multi-population and improved search operator are introduced into the CAFOABM, which can make improvement on the transition rule of search state; In addition, to avoid losing the solutions with better performance, a proportion of iterative infeasible solutions will be reserved by separating feasible solutions from infeasible solutions. Experimental results based on practical production data indicate that the zinc powder dosages which is optimized by CAFOABM are reduced approximately 7:83% comparing with the empirical solution, and the cobalt ion concentration after purification satisfies the requirement of actual production smoothly.