Cost models in nonhomogeneous Markov systems

We study the problem of introducing the phases through which a nonhomogeneous Markov system passes, as time goes to infinity. For each phase an appropriate objective function is introduced and a cost problem is formulated. Appropriate input policies subject to the cost objective functions which are established on the Markov manpower system are investigated. Finally, a procedure is provided that summarises the phases, the objective functions and the control policies commencing from the initial structure that result in its stationary form.