首页> 外文会议>International conference on numerical optimization and numerical linear algebra >A globally convergent, locally superlinearly convergent algorithm for equality constrained optimization
【24h】

A globally convergent, locally superlinearly convergent algorithm for equality constrained optimization

机译:全局会聚,局部超连续的平等约束优化收敛算法

获取原文

摘要

A new algorithm for equality constrained optimization is proposed, which solves a simple QP subproblem and an unconstrained piecewise quadratic subproblem at each iterate. The algorithm is a modification to the SQP method, but it can imporve deficiency of inconsistency of linearized constraints. Under mild condition, some global convergence results are proved. Under certain local assumptions, the algorithm generates identical iterates with SQP method. Thus, local superlinear convergence is a direct result. Some preliminary numerical results are also reported.
机译:提出了一种新的平等约束优化算法,该算法解决了一个简单的QP子问题和每个迭代的不受约束的分段二次子问题。该算法是对SQP方法的修改,但它可以减少线性化约束不一致的不足。在轻度条件下,证明了一些全球收敛结果。在某些本地假设下,算法使用SQP方法生成相同的迭代。因此,局部超连线收敛是直接结果。还报道了一些初步数值结果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号