Office Hours: By appointment (via email), Room 325, Schreiber Building
Fall semester 2024-2025: Sunday 16-19, Schreiber 007
Moed A: 16.02.25, 9 AM; Moed B: 21.03.25 9 AM
Probability Theory, Operations Research 1 (or equivalent), Introd. to Stochastic Processes (or equivalent)
Reading preparation: Yechiali, U. and Boxma, O.J.,"Poisson Processes, Ordinary and Compound", Encyclopedia of Statistics in Quality and Reliability" (F. Ruggeri, R.S. Kenett and F.W. Faltin, Eds.), Wiley, New York, 2007.
Queueing Theory deals with the analysis and design of complex stochastic systems where several 'servers' serve randomly-arriving 'customers' (e.g. jobs, calls, messages), each requiring service of random duration. While waiting for servers to become available, the customers form 'waiting lines', or queues. On the other hand, when no customers are present, the servers stay idle waiting for customers to arrive. Due to the stochastic nature of queues, neither of the two sources of idleness can be entirely eliminated. The aim of the course is to study probabilistic anlytical methods to analyse queueing systems, and to optimaly design their configurations.
Queueing theory is extensively used and applied in various areas, including Communication and Computer Networks, Service Systems, Manufacturing Systems, Road Traffic Analysis, Maintenance and Reliability, and more.
Queueing theory started in 1909 with the pioneering work of the Danish Scientist, Agner Krarup Erlang (born on January 1st, 1878), who published his paper, "The theory of probabilities and telephone conversations" (Nyt Tidsskrift for Matematik B, Vol. 20, 1909).
The course is directed to Graduate and 3rd year Undergraduate students with solid background in Probability Theory and some knowledge in Stochastic Processes.
Classification of queueing systems; The Poisson process; Exponential and Erlang (Gamma) probability distribution functions; Laplace- Stieltjes Transforms (LSTs); Probability Generating Functions (PGFs); Order Statistics; The M/G/infinity queue.
Discrete-time model; Embedded Markov chain; The Stationary distribution and its PGF; PASTA; Khinchine-Pollaczek formula; Waiting times; Little's Law; The Busy Period and its LST; Delay Busy Period; Vacation models; Polling Systems (if time permits).
The M/M/1, M/M/infinity, and M/M/c queues; Limited Buffer systems and Erlang's Loss formula; Multi-dimensional systems.
The G/M/1, G/M/c, G/Ek/1, M/D/c, and M/D/infinity queues.
Head-of-the-Line and Pre-emptive disciplines; The c-mue rule; Last-come first-served (LCFS) versus First-come first served (FCFS); Random Order of Service (ROS); Processor Sharing; Shortest Remaining Processing Time (SRPT); Right-of-Way policy.
Naor's model for M/M/1; Yechiali's model for GI/M/1 and GI/M/c.