Existence of Unique Limiting Probability Vectors in Stochastic Processes with Multiple Transition Matrices

Abstract

Concepts associated with stochastic process containing multiple transition matrices are discussed. It is proved that under certain conditions, a process with m transition matrices has m unique limiting probability vectors. This result extends the notion of discrete Markov processes to problems with intrayear and interyear dynamics. An example using a large DP model illustrates the usefulness of the concepts developed to applied problems.

Keywords

• dynamic programming
• limiting probability vectors
• stochastic processes