有约束优化预案.ppt

第四章有约束优化 North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University 总结： North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University 4.2 二次规划 二次规划：目标函数为自变量的二次函数，约束条件为线性的非线性规划问题为二次规划。 （目标函数第二项为正定二次型） North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University 例 求解二次规划 North China Electric Power University North China Electric Power University North China Electric Power University North China Electric Power University 用LP的单纯形法求解得： 原二次规划的解为： 可以验证下述条件满足K-T点： North China Electric Power University 4.3 可行方向法 North China Electric Power University 可行方向法可看作无约束下降算法的自然推广，其典型策略是从可行点出发，沿着下降的可行方向进行搜索，求出使目标函数值下降的新的可行点．算法的主要步骤是选择搜索方向和确定沿此方向移动的步长。 4.3.1 Zoutendijk 可行方向法（约坦狄克法） North China Electric Power University * North China Electric Powe