Stochastic Approximation

A class of iterative stochastic algorithms for finding the zeros or extrema of functions. In noisy environments they approach the target through recursive updates; the Robbins-Monro algorithm is their classical form.