一种求解非线性约束优化全局最优的新方法

A new method for solving global optimization solution of nonlinear unconstrained optimization problems

本文提出了一种求解非线性约束优化的全局最优的新方法—它是基于利用非线性互补函数和不断增加新的约束来重复解库恩-塔克条件的非线性方程组的新方法。因为库恩-塔克条件是非线性约束优化的必要条件,得到的解未必是非线性约束优化的全局最优解,为此,本文首次给出了通过利用该优化问题的先验知识,不断地增加约束来限制全局最优解范围的方法,一些仿真例子表明提出的方法和理论有效的,并且可行的。

In this paper, a new method is proposed for solving global optimization problem of the large-scale nonlinear constrained optimization problem, in which nonlinear equations related to Kuhn-Tucker conditions that new constrained conditions are added uninterrupted to are solved by use of nonlinear complementarily function. Because Kuhn-Tucker conditions are only necessary conditions of constrained optimization problems, a solution got by solving nonlinear functions is usual not the its global optimization solutions, for this reason, the paper given for the first time a method that add by bits constrained conditions in order to reduce feasible region of global optimization solution by use of priori information about the optimization problem. The numerical results suggest that method proposed in the paper is feasible and efficient.