Operation Research (Markov Chains) 16.3-2. Suppose that a communications network
ID: 374892 • Letter: O
Question
Operation Research (Markov Chains)
Explanation / Answer
a) We have two states, let '00' be the first and '11' the second one. Assuming such a labelling, the transition matrix is
P=(p1q1pq)P=(p1p1qq)
as, for example, the [2,1]2,1] cell represents the probability of mistake while sending '11' message. (i.e. probability of going from state '11' to state '00' in one step. Still in other words: receiving '00' when '11' is being transmitted via given link).
b) Since our Markov chain is irreducible and both states are aperiodic, the stationary distribution =[1,2]T=[1,2]T exists and is unique. We can find it by solving system of equations
[12](p1q1pq)=[12][12](p1p1qq)=[12]
together with 1+2=1.1+2=1. My computation yields 1=1q2(p+q)1=1q2(p+q) and 2=1p2(p+q)2=1p2(p+q).
c) It is the question about the probability of going from '00' to '00' in exactly two steps. Such a scenario can be realized in two ways:
Hence, the answer is p2+(1p)(1q)p2+(1p)(1q). Alternatively, you can compute the [1,1][1,1] cell of P2P2 - two-step transition probability matrix.
Ad d) One can consider 5050 as sufficiently large number of steps to approximate limnpn11limnp11n that is equal to 11 (see limiting distribution theorem). Hence, the answer is 1q2(p+q)1q2(p+q).
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.