极大熵方法与二次规划子问题的显式解

Maximum Entropy Method and Quadratic Programming SUB-problem's Explicit Solution

本文对混合约束极大极小问题的目标函数与约束分别用熵函数来逼近 ,讨论了逼近问题的二次规划子问题的搜索方向的显式形式 ,并给出了极大极小问题和多目标规划的二次规划子问题的显式解 .将所得结果用于相应的算法中 ,可提高算法的有效性 .

In this paper, the objective function and the constraints of the constrained minimax problem are approximated by the entropy functions. The explicit search directions of the Quadratic Programming (QP) sub problems of the approximating problem are discussed, and the explicit solutions of the QP sub problems of the minimax problem and the multi objective problems are obtained.The results can be applied to corresponding algorithms to improve their efficiency.