Stability and sensitivity-analysis for stochastic programming

From MaRDI portal
Publication:2277142





The survey paper presents recent developments of stability and sensitivity analysis in stochastic programming. It consists of three parts. In the introduction an exhaustive and thorough review on research directions and different approaches in stochastic programming since 1950 up to the present day is given. In the second part stability and sensitivity analysis with respect to the probability measure \(P\in {\mathcal P}\) is considered where \({\mathcal P}\) is a metrized space of probability measures. Such an approach is connected with certain results in parametric programming [see e.g. \textit{B. Bank, J. Guddat, D. Klatte, B. Kummer} and \textit{K. Tammer}, ``Nonlinear parametric optimization (Berlin 1982; for the 1983 English ed. see Zbl 0502.49002)]. The main attention is paid to quantitative stability, to Gâteaux differentiability of the optimal value function \(\phi(P)=\inf_{x\in M}f(x,P)\) and to the limit distribution of \(\phi(P_ N)\) if \(\lim_{N\to \infty}\sqrt{N}d(P_ N,P)\) is bounded in probability, where d is the distance between \(P_ N\) and P in \({\mathcal P}\). In the third part of the article the behaviour of the optimal value function and the set of optimal solutions is considered if the true probability measure P belongs to some parametric family of probability measures.



Cites work


Cited in
(54)








This page was built for publication: Stability and sensitivity-analysis for stochastic programming

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2277142)