1. Science
  2. Publications
  3. Information Processing Systems
  4. 8(145)'2016
  5. Applying the method of states parameters research to define the discrete time markov chain transitions probabilities

Applying the method of states parameters research to define the discrete time markov chain transitions probabilities

V.O. Butenko, V.Yu. Dubnitskiy, S.V. Kharchenko
Annotations languages:

The reverse problem of discrete time Markov chains analysis was defined. During the problem solution it is needed to restore the transition probabilities matrix based on the final system probabilities of being in each state. It is proposed to consider the problem as incorrect inverse problem. The smallest module method was selected as criterion of solution quality. A computational procedure for finding the solution based on a combination of research method parameter space and flexible polyhedron method. Paper contains examples of proposed procedure application.
Keywords: discrete time Markov chainsfinal states, states distribution, Markov process transition matrix, incorrect reverse problem, smallest module method, method of states parameters reserach, flexible polyhedron method