Problem 1 Denote by Pn , the long-term probability of having n people in the system Pn = lim P (X(t) = n). t→∞ Sometimes it is referred to as the steady state probabilities. For example, P6 = 0.5 implies that the probability of finding 6 people in the system is 0.5. Let’s try to derive equations for Pn under the assumption of Poisson arrival and departures with constant rate λ and µ (as you did in the project). Consider P0 . The rate at which system leaves the 0 state is λP0 (exits per minute). On the other hand, state n = 0 can only be entered by a departure from state n = 1. The rate at which system goes from state n = 1 to n = 0 is µ, and the proportion of time the system spends in state n = 1 is P1 , so the rate at which the system arrives in state n = 0 is µP1 . Since the rate at which the system enters and leaves a certain state has to be equal, we have λP0 = µP1 . Now consider state n = 1. The rate of leaving this state is (λ + µ)P1 (note the system leaves the system either by arrival or departure). The process can enter this state either by departing P0 (with rate λP0 ) or departing P2 (with the rate µP2 ). Equating rates we find (λ + µ)P1 = λP0 + µP2 . 1) Argue similarly and derive the equations for each state Pn . 2) Assume λ < µ and derive analytical expressions for all Pi , i = 0, 1, 2, ..... Note your answers will only depend on λ and µ. Plot i vs. Pi for certain large range of i (say i = 1 : 3000). 3) What will you get if you apply your results from part 2 for the case λ > µ? 4) To treat the case λ ≥ µ we will assume that the system can not have more than N customers, i.e., if the system has N customers then there is no inflow. a) Write down the equations for such system. b) Find analytical expressions for Pi , i = 1, 2, ...N . c) Plot Pi for N = 500. 1