Convergence criterion in optimization of stochastic processes

A novel approach is developed and its applicability is demonstrated for defining convergence in optimization of stochastic processes for both design and control applications. The approach is computationally simple, does not require a priori knowledge, and does not get drawn into undesirable solutions from fortuitous good values.