Страница публикации

Local search for nonsmooth DC optimization with DC equality and inequality constraints

Авторы: Strekalovsky A.S.

Журнал: Numerical Nonsmooth Optimization: State of the Art Algorithms

Том:

Номер:

Год: 2020

Отчётный год: 2020

Издательство:

Местоположение издательства:

URL:

Проекты:

DOI: 10.1007/978-3-030-34910-3_7

Аннотация: The chapter addresses the nonsmooth optimization problem with the objective function and equality and inequality constraints given by DC functions. First, the original problem is reduced to a problem without constraints by the exact penalization theory, so that the reduced (penalized) problem is also a DC minimization problem. Then, we develop a local search (LS) scheme which is based, first, on the linearization of the basic nonconvexity of the penalized problem and, second, on consecutive solutions of linearized (convex) problems. Convergence properties of the LS scheme are also investigated, which, in particular, yield that the sequence produced by LSM converges to a solution of the problem linearized at the limit point just. Moreover, the cluster point of the sequence is the KKT point for the original problem with the Lagrange multipliers provided by an auxiliary linearized problem. Finally, on the base of the developed theory several new stopping criteria are elaborated, which allow to transform the local search scheme into a local search algorithm.

Индексируется WOS: Нет

Индексируется Scopus: Нет

Индексируется УБС: Нет

Индексируется РИНЦ: Да

Индексируется ВАК: Нет

Индексируется CORE: Нет

Публикация в печати: 0