An Invariant Measure Approach to the Convergence of Stochastic Approximations with State Dependent Noise (Q3319636)

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 3850321
Language Label Description Also known as
default for all languages
No label defined
    English
    An Invariant Measure Approach to the Convergence of Stochastic Approximations with State Dependent Noise
    scientific article; zbMATH DE number 3850321

      Statements

      An Invariant Measure Approach to the Convergence of Stochastic Approximations with State Dependent Noise (English)
      0 references
      0 references
      0 references
      1984
      0 references
      invariant measure approach
      0 references
      state dependent noise
      0 references
      ordinary differential equation
      0 references
      weak convergence
      0 references
      recursive algorithms
      0 references
      adaptive control
      0 references
      The stochastic approximation process \(X_{n+1}=X_ n+a_ nf(X_ n,\xi_ n)\), \(n\geq 0\), with \(\xi_ n\) depending on \(X_ n\), is considered. Define \(t_ n=\sum^{n-1}_{i=0}a_ i\). For \(t\geq 0\), let \(x^ 0(.)\) denote the piecewise linear function with value \(X_ n\) at \(t_ n\), and let \(x^ n(t)=x^ 0(t+t_ n)\), \(t\geq 0.\) The main theorem states that \(\{x^ n(.)\), \(n\geq 0\}\) is tight and every weak limit x(.) satisfies the differential equation \(\dot x=\int f(x,\xi)P^ x(d\xi)\), where, ''loosely speaking'', \(P^ x\) is the stationary distribution of the noise \(\xi_ n\), when \(X_ n\equiv x\). The result enables one to analyse the asymptotic properties of \(X_ n\) through those of x(.). Two applications are presented.
      0 references

      Identifiers

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