Randomized algorithms for robust controller synthesis using statistical learning theory (Q4227142)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1254113
Language Label Description Also known as
default for all languages
No label defined
    English
    Randomized algorithms for robust controller synthesis using statistical learning theory
    scientific article; zbMATH DE number 1254113

      Statements

      Randomized algorithms for robust controller synthesis using statistical learning theory (English)
      0 references
      19 October 1999
      0 references
      robust controller synthesis
      0 references
      statistical learning theory
      0 references
      computational complexity
      0 references
      randomized algorithms
      0 references
      optimization
      0 references
      NP-hard problems
      0 references
      robustness analysis
      0 references
      uniform convergence of empirical means
      0 references
      0 references
      This paper is concerned with an approach to use randomized algorithms to solve several NP-hard problems in the robustness analysis and synthesis of control systems. By this approach, not ``all'' but ``most of'' the problems could be solved in polynomial-time. First, an abstract problem formulation of controller synthesis is presented and three different definitions of ``near minima'' are introduced. Then, a general approach is outlined that could be used to develop randomized algorithms for minimizing an expected controller performance defined as the objective function. Here, a property of uniform convergence of empirical means (UCEM) from statistical learning theory is shown to play a key role in finding an approximate near minimum of the objective function. Some sufficient conditions based on the notions of the Vapnik-Chervonenkis (VC-) dimension and the Pollard (P-) dimension are briefly reviewed for a family of functions to have the UCEM property. Finally, it is shown that the UCEM property holds in robust stabilization and weighted \(H_2/H_\infty\)-norm minimization problems and that randomized algorithms can solve these problems efficiently. For the entire collection see [Zbl 0901.00020].
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references