Elements of Queuing Theory: Palm-Martingale Calculus and Stochastic Recurrences (Applications of Mathematics)

This book contains an introduction to general probabilistic methods applicable to queueing systems with ergodic inputs. The main aspects treated are stability (construction of stationary states and convergence theory), time and customer averages (system equations), comparison of service disciplines and priority rules. These include, among other topics, Loynes' theory and its extensions, the basic fomulas (L=*LAMBDA*W, L=*LAMBDA*G, Kleinrock's invariance relations), PASTA, insensitivity, and optimality of SPRT. The originality of the presentation lies in the systematic use of recurrence equations to describe the dynamics of the systems, and of the Palm probability framework to describe the stationary behaviour of such systems. Also, the point of view adopted is purely probabilistic, emphasizing the use of coupling and sample path arguments (ergodic and sub-additive ergodic theory, trajectory realization of stochastic orders). The book contains an introduction on the theory of palm probability on the real line, and shows the various connections with the theory of stochastic intensity.