Enver Yücesan

Structural and behavioral equivalence of simulation models

It M sometimes desu-able to know when two different discrete-event simulation models are, in some sense, interchangeable; that is, whether or not the two models always have the same output when run under identical experimental conditions. This notion of behavioral equivalence, while conceptually simple, is difficult to define in a manner that is both useful and testable. It is difficult or impossible to assert that two simulations are behaviorally equivalent for all possible experiments. In this paper, we present an explicit and sensible defimtlon of behavioral equivalence. Unfortunately, like other definitions in the hterature, our definition is not testable in practice. However, using a general graph-theoretic specification of simulation models, which we call Simulation Graphs, we define a testable property related to behavioral equivalence which we refer to as structural equivalence. We then establish that structural equivalence (a testable property) implies behavioral equivalence (a meaningful property). This permits us to assess when it is safe to substitute one model for another. It then becomes possible to develop algorithms for addressing important problems in simulation model development and verification,

Modeling paradigms for discrete event simulation

This paper reviews modeling paradigms for discrete event simulation. Various formalisms are described and critically discussed in terms of both their ability to contribute to an improved theory of simulation and their capability for assisting in effective model construction and implementation. The mathematics of graph theory are suggested as a rigorous and comprehensive basis for specifying discrete event simulation models. In particular, Simulation Graphs are described and shown to possess many of the desirable features of a modern model-based problem-solving environment.

Complexity of simulation models: a graph theoretic approach

Complexity of a simulation model is defined as a measure that reflects the requirements imposed by models on computational resources. It is often related to the structural properties of models. In this paper, we introduce complexity measures for simulation models using the concept of Simulation Graphs. A reasonable measure of complexity is useful in a priori evaluation of proposed simulation studies that must be completed within a specified budget. They can also be useful in classifying simulation models in order to obtain a thorough test bed of models to be used in simulation methodology research. Some surrogate measures of run time complexity are also developed.

Modeling just-in-time production systems: a critical review

Kanban-controlled serial manufacturing systems have recently received considerable attention. A large proportion of the literature on the topic is devoted to success stories. There is also an important model- based effort in gaining insight into the behavior of such systems, in identifying important success factors, and ultimately in optimizing various aspects of systems performance. This paper focuses exclusively on model-based approaches in studying pull systems. Even though analytic models such as linear programming formulations or queueing approximations exist, the inherent complexity of pull systems makes simulation an essential tool in studying them. The objective of this paper is therefore to critically review selected papers that have recently appeared in refereed journals, highlight their approach, point out deficiencies, where appropriate, re-emphasize their message, and suggest new directions for research.

Teaching simulation: a panel discussion

This panel looks at the issue of teaching simulation. It brings together three individuals with a wide diversity of academic and industrial experience to discuss the key issues that should be taught in a simulation course. Questions discussed include: Should a simulation language or general modeling concepts be taught in a simulation course? Should there be a difference between simulation courses taught to engineering and business school students? What simulation tools and skills should be taught to satisfy the needs of industry who hire engineering and business school graduates? These and other issues will be discussed.

Evaluating alternative system configurations using simulation: A nonparametric approach

The real utility of simulation lies in comparing different alternatives that might represent competing system designs. Conventional statistical techniques are not directly applicable to the analysis of simulation output data in the evaluation of competing alternatives since the usual assumptions of normality and common variance are difficult to justify in simulation experiments. This paper revisits a known nonparametric test whose application has recently become feasible due to considerable increases in computing power:randomization tests assess the significance of the observed value of the test statistic by evaluating different permutations of the data. The procedure only requires invariance of the data under all permutations.

Analysis of Markov chains using simulation graph models

The construction of simulation graph models of Markov chains is demonstrated. This approach enables one to evaluate the chain either numerically through simulation or analytically through path analysis. The user can either directly simulate the associated stochastic process and obtain estimates of the desired measures of performance, or analyze the directed paths in the simulation graph model and analytically compute the probabilities of possible realizations. This method can be used in conjunction with simulation to address such problems as rare event estimation, initialization bias, and determination of initial conditions. >The construction of simulation graph models of Markov chains is demonstrated. This approach enables one to evaluate the chain either numerically through simulation or analytically through path analysis. The user can either directly simulate the associated stochastic process and obtain estimates of the desired measures of performance, or analyze the directed paths in the simulation graph model and analytically compute the probabilities of possible realizations. This method can be used in conjunction with simulation to address such problems as rare event estimation, initialization bias, and determination of initial conditions.<<ETX>>

Nonparametric techniques in simulation analysis: a tutorial

Techniques that make a minimum of assumptions about the underlying characteristics of a simulation output series are particularly useful for simulation analysis. This tutorial discusses robust nonparametric techniques with immediate applicability to such crucial steps in simulation analysis as sampling, experimental design and output analysis. Algorithms are provided for various tasks.Techniques that make a minimum of assumptions about the underlying characteristics of a simulation output series are particularly useful for simulation analysis. This tutorial discusses robust nonparametric techniques with immediate applicability to such crucial steps in simulation analysis as sampling, experimental design and output analysis. Algorithms are provided for various tasks.

Using nonparametric statistics in simulation analysis: a review

Techniques that make the minimum of assumptions about the underlying characteristics of the simulation output series are particularly useful for simulation analysis. This tutorial discusses robust non-parametric techniques with immediate applicability to such crucial steps in simulation analysis as sampling, experimental design, and output analysis. Algorithms are provided for various tasks.