一种求解混合整数非线性规划问题的演化算法--搜索空间自动收缩法

提出一种求解混合整数非线性规划问题的新的演化算法 -搜索空间自动收缩法 (ACSSOS) .在这种算法中 ,演化算法既用来定位最优解区域 ,实现搜索空间自动向全局最优解收缩 ,又用来最终求得最优解 .由于在遗传算子中引用了舍入操作 ,它不仅可用来求解混合非线性整数规划问题 ,也可求解纯整型或纯实型变量非线性函数优化问题 .数值试验结果表明本文的算法在解的质量、稳定性和收敛速度等方面优于一般的演化算法 .

A New Method of Evolutionary Algoriths for Mixed-Integer Nonlinear Optimization Problem--Automatically Constracting Search Space to Optimal Solution

A new method of evolutionary algorithms,called automatically constracting search space to optimal solution( ACSSOS) ,is proposed in this study .In this method,search space was gradually constracted to optimal solution while evolutionary algorithm was used repeatedly and incompletely.When the constracted space would cause the evolutionary algorithm trapped local optimum,the algorithm ran completely. In the evolutionary algorithm integer variables,relaxed as real variables during crossover operating,and the relaxed variables in the offsprings were rounded back before evaluating fitness. Real coding was used,and only crossoveroperators were adopted buttwo kindsof crossoveroperators were used in the method.The new method may not only be used to solve mixed- integer nonlinear optimization problems,but also used to solve the real or integer nonlinear optimization problems.The computation results demonstratored that the ACSSOS is superior to other methods in terms of solution quality , robustness and convergent speed.