Directional derivative estimates for the optimal value function of a quasidifferentiable programming problem (Q1181901): Difference between revisions

From MaRDI portal
Importer (talk | contribs)
Created a new Item
 
Added link to MaRDI item.
links / mardi / namelinks / mardi / name
 

Revision as of 23:56, 29 January 2024

scientific article
Language Label Description Also known as
English
Directional derivative estimates for the optimal value function of a quasidifferentiable programming problem
scientific article

    Statements

    Directional derivative estimates for the optimal value function of a quasidifferentiable programming problem (English)
    0 references
    0 references
    27 June 1992
    0 references
    The problem considered is of the form (P) \(f(x,y)\to\inf\), subject to \(g(x,y)\leq 0\) where \(f: \mathbb{R}^ n\times\mathbb{R}^ m\to\mathbb{R}'\) and \(g: \mathbb{R}^ n\times\mathbb{R}^ m\to\mathbb{R}'\) are quasidifferentiable, (precise definitions are given in Section 3 of the paper). Then, if \(p(y)=\inf\{f(x,y): g(x,y)\leq 0\}\) has a nonempty domain, the main results are estimates of the upper Dini directional derivative \(p^ +(y,r)\) of the optimal value function \(p\) at a fixed point \(y\) in the direction \(r\) where \(p^ +(y,r)=\limsup_{t\downarrow 0}[(p(y+tr)- p(y))/t]\). These estimates give directions of descent, or the result that the point \(y\) is stationary in ``some sense''. The main idea is to solve the substitute problem \(p(y)\to\inf\) instead of (P). An additional result is a generalization of the conclusions of the author's dissertation [Technische Universität, Chemnitz (1987)]. A list of 23 references closes the paper.
    0 references
    quasidifferentiable programming
    0 references
    nonlinear nonsmooth optimization
    0 references
    stability analysis
    0 references
    decomposition
    0 references
    upper Dini directional derivative
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references