研究了一类由连续可微函数构成的无约束Min—Max—Min问题的数值求解方法,且通过构造目标函数的极大熵函数将其转化为无约束优化问题,并建立了基本算法,给出了数值算例,表明算法是可靠和有效的.
In this paper, a numerical is proposed for solving a class of unconstrained Min - Max - Min problems, in which the objective functions are continuously differentiable. The maximum-entropy functions are used to transform the unconstrained Min- Max- Min problem into differentiable problem. A basic algorithm is established and some numerical examples are given, which prove to be reliable and effective.