混合约束非线性最优化问题的消元随机方向法

The Elimination-Random Direction Algorithm of the Nonlinear Optimization with the Mixed Restrictions

针对这样一类混合约束非线性最优化问题,即目标函数除了随设计变量变化外、还沿着时间坐标t(或某一角度坐标)变化、目标函数随时间(或角度)的变化曲线及其极值又因不同设计变量组而异、最优设计应使目标函数随时间(或角度)变化的最大值为最小的最优化问题,提出了一种称为消元随机方向搜索法的新算法,介绍了这一算法的基本思想,给出了该算法的迭代计算程序框图.

Aiming at such a type of the mixed restriction nonlinear optimization issues, we know that the target function varies with the time coordinates （or a certain angle coordinates） in addition to its change with the design variables set, that the curve and the maximum value of the target function varying with time（or angle） is different if the design variables set is different, and that the optimal design variables set should make the maximum value of the target function varying with time （or angle） minimum. A new algorithm referred to as the elimination-random direction algorithm is put forward and its basic principle is explained, and an iterative computation flow and a procedure chart are presented.