Queueing Systems

Queueing Systems
Name	Queueing Systems
Discipline	Operations research, Applied mathematics

Queueing Systems Queueing systems are mathematical frameworks that model the behavior of waiting lines and service processes used to analyze capacity, delay, and resource allocation. Scholars and practitioners from Agner Krarup Erlang's time to modern researchers at Bell Labs, MIT, Stanford University, Princeton University, University of Cambridge and University of Oxford have advanced theory and applications across telecommunications, transportation, healthcare, and computing. Foundational results connect to work by Andrey Kolmogorov, Albert Einstein, Harry Nyquist, Norbert Wiener and methods used in Winston Churchill-era logistics, while contemporary developments intersect with research at Google, Amazon (company), Microsoft, IBM, and Intel Corporation.

Introduction

Queueing systems study stochastic processes that represent arrivals, services, and departures in systems such as call centers at AT&T, packet switches in Cisco Systems routers, checkout lines in Walmart stores, and patient flows in Mayo Clinic. Early models drew on empirical observations from telephony at Kongelige Telegrafverk and experimental work by Erlang for the Copenhagen telephone system. Modern theory synthesizes contributions from mathematicians and engineers affiliated with institutions like Courant Institute, Institut Henri Poincaré, and Max Planck Society and shapes policy and design at organizations including Federal Communications Commission and European Telecommunications Standards Institute.

Mathematical Models and Notation

The canonical notation for single- and multi-server models often uses the Kendall compact form introduced by David George Kendall, with variants studied by researchers at Columbia University and University of California, Berkeley. Key stochastic processes include Poisson processes related to Srinivasa Ramanujan-type distributions, renewal processes influenced by work at Bell Labs, birth–death processes studied by Franz Ernst Neumann-influenced schools, and Markov chains developed in the tradition of Andrey Kolmogorov and Émile Borel. Notation commonly specifies arrival distributions, service distributions, number of servers, population size, and queue capacity; seminal textbooks from John Kingman, Leonard Kleinrock, and Donald Gross codified these conventions. Advanced models integrate Lévy processes analyzed by scholars at Courant Institute and point processes used in research at Imperial College London.

Performance Measures and Analysis Techniques

Performance measures include average waiting time explored in studies by Agner Krarup Erlang and David George Kendall, queue length examined in work at Princeton University, sojourn time characterized in research from Stanford University, and throughput metrics applied at Bell Labs and AT&T. Analytical techniques employ generating functions developed in the tradition of G. H. Hardy and John Littlewood, transform methods traceable to Carl Friedrich Gauss-linked analysis, matrix-analytic methods advanced by groups at Queen's University, heavy-traffic approximations connected to the Central Limit Theorem lineage, and Palm calculus rooted in research at École Polytechnique. Stability criteria and ergodicity results reference contributions by Kolmogorov and Andrei Nikolaevich Kolmogorov-inspired schools, while recent probabilistic coupling and sample-path large deviations draw on work associated with Ludwig Boltzmann-influenced statistical mechanics.

Common Queueing Disciplines and Variants

Service disciplines include first-come, first-served used in operational settings at Walmart and McDonald's, last-come, first-served explored in theoretical papers from University of Chicago, processor sharing motivated by time-sharing systems at MIT, priority queues central to scheduling in NASA missions, and round-robin influenced by designs at Intel Corporation. Variants cover birth–death queues, batch arrivals studied at Bell Labs, networks of queues popularized by Jackson network-type analyses from James R. Jackson's era, closed-network models used in manufacturing plants at Toyota, loss systems like Erlang B applied to trunking in British Post Office history, and retrial queues researched at Moscow State University.

Applications and Case Studies

Applications range from telephone traffic engineering in early work at Erlang's Copenhagen office and AT&T research labs to packet switching designs at Xerox PARC and Cisco Systems. Transportation deployments involve scheduling at Heathrow Airport and dispatching studies for Union Pacific Railroad; healthcare implementations include patient flow optimization at Mayo Clinic and ambulance routing models used by London Ambulance Service. In computing, load balancing at Google and caching strategies at Amazon (company) draw on queueing theory. Case studies of large-scale systems cite empirical analyses from Bell Labs, field trials commissioned by Federal Aviation Administration, and capacity planning at Goldman Sachs.

Simulation and Numerical Methods

Monte Carlo simulation techniques used in studies at Los Alamos National Laboratory and Sandia National Laboratories complement numerical solutions such as matrix-geometric methods developed at McGill University and approximations derived from heavy-traffic limits associated with Stanford University. Discrete-event simulation packages originating from projects at IBM and University of Waterloo enable emulation of complex service networks. Numerical inversion of transforms, finite-state truncation, and iterative solvers trace intellectual roots to computational mathematics groups at Institute for Advanced Study and National Institute of Standards and Technology.

History and Key Contributions

The field originated with Agner Krarup Erlang's telephony analyses in Copenhagen and expanded through contributions by David George Kendall, John Littlewood, Leonard Kleinrock's internet-era investigations, and John Kingman's foundational asymptotic results. Other notable contributors include Dennis G. Kendall-era collaborators, scholars at Bell Labs such as William Feller-influenced probabilists, and contemporary researchers across MIT, Princeton University, UC Berkeley, and INRIA who advanced stochastic networks and performance evaluation. Institutional milestones include the growth of operations research departments at INSEAD, London School of Economics, and the establishment of conferences sponsored by Institute for Operations Research and the Management Sciences.

Category:Operations research