局部搜索微分进化算法求解复杂通风网络优化策略研究

A Differential Evolution Algorithm with Local Search Strategy for Solving the Optimal Problem of Complex Ventilation Network

摘要: 针对微分进化算法局部收敛性较差的缺点，提出了种群分类进化方法，将种群个体划分为最优矢量，比较优矢量和较差矢量，在每一代种群进化过程中，保留最优矢量；在较优矢量邻域范围内搜索更优矢量替代当前矢量；重新生成优于较差矢量新矢量替换原来矢量。建立了矿井通风网络优化模型，提出了模型的降维方法和线性化转化方法，引入局部搜索策略，通过对父代个体邻域进行深度搜索，选择不劣于父代的个体组成子代，提出了局部搜索微分进化算法，与关键路径法相结合，实现了对矿井通风网络优化调节模型的求解。通过单风机和多风机复杂通风网络实例，验证了算法的可行性，获得了矿井通风网络优化的最优调节方案。

Abstract: A population classification evolution method is proposed to solve the problem of poor local convergence of differential evolution algorithm. The population is divided into the optimal vector, the superior vector and the inferior vector. In each generation of population evolution, the optimal vector is retained. Search for a better vector to replace the current vector in the neighborhood range of the superior vector. Replace the original vector with a new vector that is superior to the inferior vector. The optimization model of mine ventilation network is established, the dimensionality reduction method and linearization transformation method of the model are put forward, and local search strategy is introduced. Through the deep search of the neighborhood of the individual in the parent generation, the individual that is not inferior to the parent generation is selected to constitute the child generation. The local search differential evolution algorithm is put forward, combined with the critical path method, to realize the solution of the mine ventilation network optimization regulation model. An example of complex ventilation network with single fan and multifan is given to verify the feasibility of the algorithm. The optimal regulation scheme of mine ventilation network optimization is obtained.