Discrete-event simulation. Modeling, programming, and analysis

Publication date 30 August 2001

Published in Springer Series in Operations Research (Search for Journal in Brave)

complexity discrete-event systems modelling simulation

Probabilistic models, generic numerical methods in probability and statistics (65C20) Introductory exposition (textbooks, tutorial papers, etc.) pertaining to computer science (68-01) Introductory exposition (textbooks, tutorial papers, etc.) pertaining to systems and control theory (93-01) Discrete event control/observation systems (93C65)

Review text

The first step in studying a system consists in converting accumulated knowledge into a model, i.e., into a representation based on theory and/or previously, to some extent, structured empirical observations. Being already available, it is expected that by using the model one can gain more theoretical and experimental knowledge about the original system. According to the complexity of the model one approaches it analytically, by simulation or by combining these two procedures. In this respect and in many cases that are very relevant for applications, the discrete-event systems and their apparent models exhibit a great complexity. In order to properly deal with them, one mostly proceeds by simulation because the values of system performance descriptors are unknown, and determining them remains beyond the capacity of the analytical tools of queueing theory. Discrete-event simulation offers techniques that can approximate the values of these descriptors to within a remarkably small error. The approximations come from analyses of data observed on sample paths generated during execution of a simulation program corresponding to the given model, usually on adequate queue-length and waiting-time processes. At the same time, once a simulation procedure is adopted, one must apply the simulation techniques interactively with the adaptation of the modelling techniques because the designer repeatedly faces the problem of balancing the need for structural detail with the need to make the model amenable to problem solving techniques. In this context the book is a valuable basis for acquiring the necessary skills by describing in nine chapters concepts, methods and providing guidance for: -- modelling the behaviour of a discrete-event system, translating the model into code executable on a computer; -- executing the code and generating output in response to one or more input scenarios, some assigning different numerical values to system parameters and others based on different strategies for the dynamically operating system; -- analysing each output to infer the system behaviour that each scenarios induces; this is done by assessing, by means of statistical methods, how well estimates of system performance measures approximate their true, but unknown, values; -- comparing these inferred behavioural patterns for different scenarios; -- computational complexity as a function of problem size; -- sampling from diverse distributions on a computer; -- pseudorandom number generation; -- using empirical data to estimate the values of the input parameters of a simulation model. The book is highly recommended as an introductory course for graduate and advanced undergraduate students in the mathematical and engineering sciences, particularly operations research, industrial engineering, operations management, computer science, telecommunications engineering, transportation engineering, and to the professionals in these fields who are interested to broaden their knowledge of discrete-event simulation.

Recommendations

Cited in

(89)

This page was built for publication: Discrete-event simulation. Modeling, programming, and analysis

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2737800)