Hook-Jeeves方法在线性不等式约束优化中的推广

Extentions of Hook-Jeeves Method to Optimization with Linear Inequality Constraints

分别将无约束优化的线搜索和离散步Hook-Jeeves算法推广到带一般线性不等式的约束优化,并产生两个可行下降算法;在无任何约束规格的情况下,证明了线搜索的Hook-Jeeves算法具有全局收敛性.

The algorithm of Hook-Jeeves using line searches and the algorithm of Hook-Jeeves with discrete sreps are extended such that they can solve optimization problems with linear inquality constraints. We proposed two improving feasible algorithms. Under no constraint qualification conditions, we prove that the extended algorithm of Hook-Jeeves with linesearches prossesses global convergence.