Relaxation and Regularization of Nonconvex Variational Problems
نویسندگان
چکیده
We are interested in variational problems of the form min ∫ W (∇u) dx, with W nonconvex. The theory of relaxation allows one to calculate the minimum value, but it does not determine a well-defined “solution” since minimizing sequences are far from unique. A natural idea for determining a solution is regularization, i.e. the addition of a higher order term such as ǫ|∇∇u|2. But what is the behavior of the regularized solution in the limit as ǫ → 0? Little is known in general. Our recent work [19, 20, 21] discusses a particular problem of this type, namely minuy=±1 ∫ ∫ ux + ǫ|uyy| dxdy with various boundary conditions. The present paper gives an expository overview of our methods and results. Partially supported by NSF grant DMS-9102829, AFOSR grant 90-0090, and ARO contract DAAL0392-G-0011. Partially supported by NSF grant DMS-9002679 and by SFB 256 at the University of Bonn.
منابع مشابه
Optimality conditions for approximate solutions of vector optimization problems with variable ordering structures
We consider nonconvex vector optimization problems with variable ordering structures in Banach spaces. Under certain boundedness and continuity properties we present necessary conditions for approximate solutions of these problems. Using a generic approach to subdifferentials we derive necessary conditions for approximate minimizers and approximately minimal solutions of vector optimizatio...
متن کاملGlobal Solutions of Variational Models with Convex Regularization
We propose an algorithmic framework for computing global solutions of variational models with convex regularity terms that permit quite arbitrary data terms. While the minimization of variational problems with convex data and regularity terms is straightforward (using, for example, gradient descent), this is no longer trivial for functionals with nonconvex data terms. Using the theoretical fram...
متن کاملNonconvex Evolution Inclusions Generated by Time - Dependent Subdifferential Operators
We consider nonlinear nonconvex evolution inclusions driven by time-varying subdifferentials 0(t,x) without assuming that (t,.) is of compact type. We show the existence of extremal solutions and then we prove a strong relaxation theorem. Moreover,r we show that under a Lipschitz condition on the orientor field, the solution set of the nonconvex problem is path-connected in C(T,H). These result...
متن کاملInterior-point methods for nonconvex nonlinear programming: regularization and warmstarts
In this paper, we investigate the use of an exact primal-dual penalty approach within the framework of an interior-point method for nonconvex nonlinear programming. This approach provides regularization and relaxation, which can aid in solving ill-behaved problems and in warmstarting the algorithm. We present details of our implementation within the loqo algorithm and provide extensive numerica...
متن کاملGlobally convergent edge-preserving regularized reconstruction: an application to limited-angle tomography
We introduce a generalization of a deterministic relaxation algorithm for edge-preserving regularization in linear inverse problems. This algorithm transforms the original (possibly nonconvex) optimization problem into a sequence of quadratic optimization problems, and has been shown to converge under certain conditions when the original cost functional being minimized is strictly convex. We pr...
متن کامل