Convergence Theorems for a Class of Recursive Stochastic Algorithms

摘要: Several recent studies of the way individual economic units might learn their parts in an or strategic equilibrium have modeled learning process as a recursive algorithm with stochastic features. The central idea each these has been to explain justify notion by demonstrating that is stationary point which converges. This approach analysis—and particular modelling terms algorithm—seems attractive and powerful, we expect its use become more widespread.