Convergence of a random algorithm for function optimization
C. C. Y. Dorea, Liang Zhao · Numerical Functional Analysis and Optimization · 1999
Consider a non-homogeneous Markov chain {Xn } n ≥ 0 to be used to estimate the global minimum of a given function f defined on S = {1, 2, …N}. We show how ergodicity results can be successfully applied to obtain the desired convergence where