期刊
SIAM JOURNAL ON OPTIMIZATION
卷 18, 期 4, 页码 1187-1208出版社
SIAM PUBLICATIONS
DOI: 10.1137/060657741
关键词
semi-infinite programming; alpha BB; global optimization; convex optimization; mathematical program with complementarity constraints; bilevel optimization
资金
- Deutsche Forschungsgemeinschaft
We present a new numerical solution method for semi-infinite optimization problems. Its main idea is to adaptively construct convex relaxations of the lower level problem, replace the relaxed lower level problems equivalently by their Karush-Kuhn-Tucker conditions, and solve the resulting mathematical programs with complementarity constraints. This approximation produces feasible iterates for the original problem. The convex relaxations are constructed with ideas from the alpha BB method of global optimization. The necessary upper bounds for second derivatives of functions on box domains can be determined using the techniques of interval arithmetic, where our algorithm already works if only one such bound is available for the problem. We show convergence of stationary points of the approximating problems to a stationary point of the original semi-infinite problem within arbitrarily given tolerances. Numerical examples from Chebyshev approximation and design centering illustrate the performance of the method.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据