组织进化数值优化算法

基于经济学中“组织”的概念 ,该文提出一种新的进化算法———组织进化算法 ,来解决无约束和有约束的数值优化问题 .该算法与传统遗传算法、进化规划、进化策略的运行机制完全不同 ,其进化操作不直接作用于个体上 ,而作用在组织上 ,为此 ,该文定义了三种组织进化算子———分裂算子、吞并算子和合作算子来引导种群进化 .理论分析证明组织进化算法具有全局收敛性 .实验中 ,用 4个无约束和 6个有约束标准函数对算法进行了测试 ,与 3个新算法作了比较 ,并对组织进化算法的性能作了深入分析 .结果表明 ,该文算法无论在解的质量上还是在计算复杂度上都优于其它算法 .对于有约束问题 ,只用了简单的静态罚函数就得到了良好的效果 ,这表明该文算法的搜索机制非常有效 ,不易陷入局部最优 .最后 ,参数分析的结果表明该文算法具有性能稳定、成功率高、对参数不敏感等优越的性能

An Organizational Evolutionary Algorithm for Constrained and Unconstrained Optimization Problems

Based on the concept of organization in economics, a novel evolutionary algorithm, Organizational Evolutionary Algorithm (OEA), is proposed to deal with both unconstrained and constrained optimization problems. Its mechanism is completely different from the traditional genetic algorithm, evolutionary programming and evolution strategy. In OEA, evolutionary operations do not act on individuals directly, but on organizations, so three evolutionary operators is designed for organizations. Theoretical analysis proves that OEA converges to the global optimum. In experiments, OEA is tested on 4 unconstrained and 6 constrained benchmark problems, and compared with three recent algorithms. The results indicate that OEA performs much better than the three other algorithms both in the quality of solution and in the computational complexity. For constrained problems, good results are obtained by only using a very simple penalty term, which shows that OEA has a low probability to get trapped in the local optima and its search mechanism is very effective. In addition, parameter analysis demonstrate that OEA has stable performance and high success ratio, and is insensitive to parameters.