Chapter 2 Machine Interference Model Long Run Analysis Deterministic Model Markov Model Problem Description • • • • Group of m automatic machines Operator must change tools or perform minor repairs How many machines should be assigned to one operator? Performance measures – Operator utilization: = fraction of time the operator is busy – Production rate: TH = # finished items per unit time – Machine availability: = TH/G, where G is the gross production rate, or the production rate that would be achieved if each machine were always available • Note: In this queuing system, the machines are the customers! IE 512 Chapter 2 2 Long Run Analysis Each machine has gross production rate h Pn is the proportion of time that exactly n machines are down: m n 0 Pn 1 Then, given Pn, TH n 0 m n hPn m m nP TH n 0 m n hPn 1 n mh mh n 0 m 1 P0 m IE 512 Chapter 2 3 Eliminate some unknowns Suppose the mean time to repair a machine is 1/, and the mean time between failures for a single machine is 1/. D t = avg. # of repairs in (0,t] = t m A t = avg. # of failures in (0,t] = n0 m n Pnt mt In the long run, assuming the system is stable, D t A t , so that t mt , and m h TH mh IE 512 1 P0 is still unknown... Chapter 2 4 Queuing Measures of Performance • L = average # of machines waiting for service n 1 P nP 1 P m m n 1 n 1 n m 1 • m n 1 nPn m 1 T = average downtime duration of a machine m Total machine-hrs down in (0,t] n1 nPnt m 1 t 1 A t • 0 N = average number of machines down • n mt mt W = average duration of waiting time for repair T IE 512 1 1 1 Chapter 2 5 Little’s Formula Observe from the previous equations: N T m 1 1 L m m W where m is the total average number of failures per unit time = the arrival rate of customers to the queuing system Little’s formula relates mean # of customers in system to mean time a customer spends in the system. IE 512 Chapter 2 6 A Deterministic Model Suppose each machine spends exactly 1/ time units working followed by exactly 1/ time units in repair. Then if m 1 1 1 and we could stagger the failure times, we would have no more than one machine unavailable at any time, so that m 1 m h m h 1 1 , , TH m (Otherwise, IE 512 1, h , TH m Chapter 2 7 A Markov Model j exp be the time between the (n-1)st repair Let n and the nth failure of machine j, and Sn exp be the time duration of the nth repair (indep.) The time until the first failure is min 11 ,1 2 ,...,1 m exp m N(t) = # of machines down at time t follows a CTMC with S = {0, 1, …, m} and m Q 0 0 IE 512 m m 1 0 m 1 0 0 0 0 m n m n 0 0 Chapter 2 0 0 0 8 Steady-State Probabilities pn lim Pr N t n t satisfy the balance equations or level-crossing equations m p0 p1 m p0 p1 m 1 p 1 m 1 p1 p2 m p0 p2 m n p m n 1 p n n 1 pn 1 pm 1 pm pm pm1 IE 512 m n pn pn1 Chapter 2 9 Solution n m! pn p0 and m n ! Say ˆ m! n0 pn 1 p0 m n ! n 0 m m n 1 1 k k! and let G ˆ , k n0 ˆ n , so that p0 G ˆ , m k n ! Then G ˆ , k 1 k ˆ G ˆ , k 1 , k 0,1,... G ˆ , m 1 1 1 1 m m ˆ m ˆ G ˆ , m G ˆ , m IE 512 Chapter 2 10 Erlang Distribution If failure and/or repair times are not exponential, can fit an Erlang distribution by matching moments: Solve X k , S 2 k 2 simultaneously for k and Big advantage: Can still model as a CTMC. Consider time to machine failure (each machine) as Erlangk. Can think in terms of k phases in the time to failure, where the time the m/c spends in each phase is exponential (k): 1 Mean time spent in each phase = k 1 1 k Mean total time to failure = k IE 512 Chapter 2 11 Expanded State Definition Mi(t) = # of machines operating in phase i at time t (so N t m i 1 M i t ) k and M 1 t , M 2 t ,..., M k t is a CTMC with S l1 , l2 ,..., lk , 0 li m, i 1,..., k and 0 i 1 li m k For example, if k = 2, then a single machine without interference follows the CTMC (1 = failed state): 0;1 2 IE 512 0;2 Chapter 2 2 1 12 Transitions among States (k=2) Steady state probabilities: p l1 , l2 Pr li machines are operating in phase i, i 1, 2 Say l l1 l2: Rate out of state l1 , l2 is 2 l1 l2 2l Rate into state l1 1, l2 2(l1+1) l1 , l2 2(l2+1) IE 512 l1 1, l2 1 Chapter 2 l1, l2 1 13 Balance Equations l 0 1 l m 1 lm p 0, 0 2 p 0,1 2l p l1 , l2 p l1 1, l2 2 l1 1 p l1 1, l2 1 2 l2 1 p l1 , l2 1 2 mp l1 , l2 p l1 1, l2 2 l1 1 p l1 1, l2 1 This system of equations (for any k) has the solution: p l1 , l2 ,..., lk IE 512 l k l1 !l2 ! lk ! Chapter 2 p 0, 0,...0 14 SS Number of Machines Working From the previous equation and p l l1 l2 p l , l ,..., l 1 2 k lk get p l Pm l (probability that l machines are working) = p 0 l! l (Note: typo in (2.47), p.33 text) Find probabilities by normalizing to 1. This distribution is independent of k or any other characteristics of the failure time distribution. It can be shown that the same state distribution holds for any failure time distribution! IE 512 Chapter 2 15