摘要
本文研究求解不等式约束离散ｍｉｎｉｍａｘ问题的区间算法，其中目标函数和约束函数是 Ｃ~１类函数．利用罚函数法和极大熵函数思想将问题转化为无约束可微优化问题，讨论了极大熵函数的区间扩张，证明了收敛性等性质，提出了无解区域删除原则，建立了区间极大熵算法，并给出了数值算例．该算法是收敛、可靠和有效的．
An interval algorithm for inequality constrained discrete minimax problems is described,in which the constituent objective functions and constrained functions are C^1 functions. We transform this problem to unconstrained differentiable optimization problem with the idea of the maximum-entropy function and penalty function methods, discuss the interval extensions of the maximum-entropy functions, prove relevant properties, and provide the region deletion test rules. At last, we design an interval maximum-entropy algorithm with the bisection rule of Moore. That is convergence, stable and reliable. Moreover, the numerical results are presented.
出处
《运筹学学报》
CSCD
1999年第4期55-64,共10页
Operations Research Transactions
基金
The project was supported by the Coal Science Foundation of China ,the Science Foundation of CUMT
关键词
区间算法
不可微优化
区间极大熵法
不等式约束
interval algorithm, maximum-entropy function, penalty function, discrete minimax problem.
