提出了一个基于降维技术的填充函数方法,用以求解箱约束非线性全局优化问题。首先利用降维变换将n维问题转化为一维问题,其次对一维问题运用填充函数方法求解,证明了降维填充函数的理论性质,并给出了算法和数值实验结果。
This paper presents a filled function method based on reducing dimension technology. The method will be used for the nonlinear global optimization problems with box constraints. Firstly, a reducing transformation is used to convert an n-variable problem into a one-variable problem. Secondly, the one-variable problem is solved by filled function method. The paper proves the theoretical characteristic of filled function, gives the algorithm and lists the experimental results at last.