CS2-04: Time-homogeneous Markov jump processes 6 AGOGO CDO Page 27 The Poisson process revisited We have already mentioned (at the end of Section 2) that the Poisson process can be formulated as a Markov jump process. We now revisit this idea. Consider the Markov jump process with state space S = {0, 1, 2, ...} and transition rates: if j = i if j = i 1 otherwise ij 0 The diagram representation is: 0 1 2 i 3 i +1 Recall that, in a Poisson process, events occur one at a time, and it is impossible to move to a lower-numbered state. So, when it leaves state i , the process must enter state i 1 . The generator matrix A in Kolmogorov’s equations is: 0 A 0 This leads to the forward equations: pi 0 (t ) pij (t ) pi 0 (t ) pij 1(t ) pij (t ), j 0 essentially identical to (4.3) and (4.4). It is interesting also to consider the backward equations: pij (t ) pij (t ) pi 1, j (t ) which of course have the same solution as the forward equations despite looking dissimilar. The Actuarial Education Company © IFE: 2019 Examinations