非单调线搜索解线性约束优化的信赖域内点算法

A Nonmonotone Trust Region Interior Algorithm for Linearly Constrained Optimization

提供了求解线性约束的非线性优化问题的非单调信赖域内点算法.在合理的条件下,证明了算法的整体收敛性,并且在最优解局部范围内获得单位步长的可接受性,从而保证了局部超线性收敛速率.

We present a trust region interioralgorithm in association with nonmonotone line search technique for linearly constrained optimization. We prove that, under suitable hypotheses, the proposed algorithm converges globally, and analyze the unit stepsize will be asymptotically accepted.