ANALYSIS OF CONVERGENCE PROPERTIES OF A STOCHASTIC-EVOLUTION ALGORITHM

In this paper, the convergence properties of a stochastic optimization algorithm called the stochastic evolution (SE) algorithm is analyzed, We show that a generic formulation of the SE algorithm can be modeled by an ergodic Markov chain, As such, the global convergence of the SE algorithm is established as the state transition from any initial sta te to the globally optimal states, We propose a new criterion called t he mean first visit time (MFVT) to characterize the convergence rate o f the SE algorithm, With MFVT, we are able to show analytically that o n average, the SE algorithm converges faster than the random search me thod to the globally optimal states, This result is further confirmed using the Monte Carlo simulation.