非线性0-1规划问题的元胞灰狼优化算法

Solving Nonlinear 0-1 Programming Problems by Cellular Gray Wolf Optimization Algorithm

针对非线性0-1规划求解问题,基于元胞自动机原理和改进的灰狼算法,提出一种元胞灰狼优化算法.首先,为了避免基本灰狼算法种群分布的随机性问题,利用佳点集理论对灰狼种群进行初始化,增强算法种群的多样性,提高算法的全局收敛速度;其次,针对基本灰狼算法的开发和探索能力平衡能力差的问题,利用自适应精英学习策略分别对算法中的参数α、灰狼与猎物的距离进行修正,实现灰狼算法的全局搜索速度和开发探索能力的最优均衡性;最后,将元胞自动机的演化规则与次优解β灰狼位置以及第三优解δ灰狼位置进行更新,利用元胞及其邻居增强搜索过程的多样性和分布性,实现灰狼算法的全局优化能力;并选用14个典型的非线性0-1规划问题算例进行仿真解算,并将解算结果与其它算法进行比较,结果表明:该算法对大规模复杂问题求解的平均运行时间少10%左右,且具有较快的收敛速度、较多的最优解集和较好的全局寻优能力.

According to the problem of non-linear 0-1 programming,a cellular gray wolf optimization algorithm has proposed,which was cellular automata and improved gray wolf algorithm.Firstly,for initialize the gray wolf population through use the good point set theory,in order to avoid the randomness of the population distribution,enhance the diversity of the population and improve the global convergence speed of the algorithm.Secondly,Adaptive elite learning strategies was used to modify the parameters of the algorithm,the distance between the gray wolf and the prey,for balancing development and exploration ability of gray algorithm.Finally,the cell and its neighbors utilized to update the position of the second-best solution gray wolf and the third-best solution gray wolf,solving the global optimization of the gray wolf algorithm.Taking fourteen typical examples of nonlinear 0-1 programming problems as example,and compared with other algorithms.The results show that the average running time of the algorithm is less than 10%,and has a faster convergence speed,more optimal solution sets,and better global optimization ability.