离散全局最优化中的一类T-F函数算法(英文)

Class of T-F Functions for Discrete Global Optimization

针对求解非线性离散规划全局最优解问题提出一类T-F函数算法。首先,介绍有关离散全局最优解的各种概念,并定义了T-F函数;其次,提出一类T-F函数,并设计了相应的T-F函数算法,通过寻找该T-F函数的离散局部极小解,以期找到离散规划问题的比当前离散局部极小解更好的解。数值实验表明算法是有效的。

A class of new T-F function for solving discrete general minimization problems with a general function over box-constrained domain is introduced. A T-F function is constructed at a local minimizer of the objective function such that it achieves local maximum at the current solution. Moreover, a local minimizer of the T-F function leads to a, new solution to the original problem with lower objective function value. Iteration follows in this manner to reach a global minimizer. Promising computational results are included and show the efficiency of the T-F function method.