No label defined (Q3735490)

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 3968351
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3968351

      Statements

      0 references
      0 references
      1986
      0 references
      Iterative stochastic optimization algorithms
      0 references
      estimation of the convergence rate
      0 references
      [For the entire collection see Zbl 0585.00014.] Iterative stochastic optimization algorithms are widely used to solve problems of estimation, adaptation, learning, etc. It is clearly important to study the convergence, rate of convergence, stability and other properties of such procedures. For some time, the question of convergence has received a good deal of attention while the rate of convergence has received relatively little. The few estimates of convergence rates given in the literature are generally asymptotic in character. In addition, they have been obtained only for those optimization problems which satisfy the condition of strong convexity or the condition of strong pseudo-differentiability. Again, when using iterative methods it is necessary to estimate the smallest number of iterations required to obtain a given accuracy. It is clear that asymptotic estimates of the convergence rate are not suitable for this purpose. In addition, such estimates do not generally suggest any course of action regarding the iterative process as a whole. This paper summarizes results obtained recently by the authors on the non-asymptotic (valid from the first step) analytical estimation of the convergence rate of stochastic iterative algorithms. We study both singular and non-singular problems, including those subject to the conditions of strong and uniform pseudo-differentiability.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references