A validated simulation model primarily requires performing an appropriateinput analysis mainly by determining the behavior of real-world processes usingprobability distributions. In many practical cases, probability distributionsof the random inputs vary over time in such a way that the functional forms ofthe distributions and/or their parameters depend on time. This paper answersthe question whether a sequence of observations from a process follow the samestatistical distribution, and if not, where the exact change points are, sothat observations within two consecutive change points follow the samedistribution. We propose two different methods based on likelihood ratio testand cluster analysis to detect multiple change points when observations follownon-stationary Poisson process with diverse occurrence rates over time. Resultsfrom a comprehensive Monte Carlo study indicate satisfactory performance forthe proposed methods. A well-known example is also considered to show theapplication of our findings in real world cases.
展开▼
