Identifying Active Constraints via Partial Smoothness and Prox-Regularity
نویسندگان
چکیده
Active set algorithms, such as the projected gradient method in nonlinear optimization, are designed to “identify” the active constraints of the problem in a finite number of iterations. Using the notions of “partial smoothness” and “prox-regularity” we extend work of Burke, Moré and Wright on identifiable surfaces from the convex case to a general nonsmooth setting. We further show how this setting can be used in the study of sufficient conditions for local minimizers.
منابع مشابه
Identifying Active Manifolds
Determining the “active manifold” for a minimization problem is a large step towards solving the problem. Many researchers have studied under what conditions certain algorithms identify active manifolds in a finite number of iterations. In this work we outline a unifying framework encompassing many earlier results on identification via the Subgradient (Gradient) Projection Method, Newton-like M...
متن کاملPartial second-order subdifferentials of -prox-regular functions
Although prox-regular functions in general are nonconvex, they possess properties that one would expect to find in convex or lowerC2 functions. The class of prox-regular functions covers all convex functions, lower C2 functions and strongly amenable functions. At first, these functions have been identified in finite dimension using proximal subdifferential. Then, the definition of prox-regula...
متن کاملSweeping process by prox-regular sets in Riemannian Hilbert manifolds
— In this paper, we deal with sweeping processes on (possibly infinitedimensional) Riemannian Hilbert manifolds. We extend the useful notions (proximal normal cone, prox-regularity) already defined in the setting of a Hilbert space to the framework of such manifolds. Especially we introduce the concept of local prox-regularity of a closed subset in accordance with the geometrical features of th...
متن کاملRegularization of differential variational inequalities with locally prox-regular sets
This paper studies, for a differential variational inequality involving a locally prox-regular set, a regularization process with a family of classical differential equations whose solutions converge to the solution of the differential variational inequality. The concept of local prox-regularity will be termed in a quantified way, as (r, α)-prox-regularity.
متن کاملA proximal method for identifying active manifolds
The minimization of an objective function over a constraint set can often be simplified if the “active manifold” of the constraints set can be correctly identified. In this work we present a simple subproblem, which can be used inside of any (convergent) optimization algorithm, that will identify the active manifold of a “prox-regular partly smooth” constraint set in a finite number of iterations.
متن کامل