Markov chain transition probabilities and experimental data (Q1117587)

Markov chain theory is used in a number of applications to model uncertainties in dynamic behavior. The theory is attractive because it is the probabilistic analogue to the classical physics approach to dynamics and the state transition matrix is an ideal probabilistic counterpart to the deterministic matrix concept. The determination of the transition matrix is the main problem of the method. The author, by decomposing each state into ``loading'' and ``unloading'' states, develops a useful technique by which experimental data can be used to estimate the probabilities of the transition matrix.