WSC 2008 Final Abstracts

Analytical Simulation Modeling
Lee Schruben (University of California, Berkeley)

Abstract:
Simulation modeling methodology research and simulation analysis methodology research have evolved into two nearly separate fields. In this paper, ways are shown how simulation might benefit from modeling and analysis becoming more closely integrated. The thesis of this paper is that simulation analysis and simulation modeling methodologies, considered together, will result in important advancements in both. Some examples demonstrate how dramatically more efficient discrete event simulation models can be designed for specific analytical purposes, which in turn enable more powerful analytical procedures that can exploit the special structures of these models. A series of increasingly difficult analytical problems, and models designed to solve them, are considered: starting with simple performance estimation, and progressing to dynamic multiple moment response surface meta-modeling.

The Mathematics of Continuous-Variable Simulation Optimization
Sujin Kim (National University of Singapore) and Shane G. Henderson (Cornell University)

Abstract:
Continuous-variable simulation optimization problems are those optimization problems where the objective function is computed through stochastic simulation and the decision variables are continuous. We discuss verifiable conditions under which the objective function is continuous or differentiable, and outline some key properties of two classes of methods for solving such problems, namely sample-average approximation and stochastic approximation.

Incorporating Information Networks Into Military Simulations
Darryl K. Ahner (United States Military Academy), Jonathon K. Alt and Francisco K. Baez (U.S. Army TRADOC Analysis Center), John Jackson (U.S. Joint Forces Command J9) and Thorsten Seitz and Susan M. Sanchez (Naval Postgraduate School)

Abstract:
Information superiority is considered a critical capability for future joint forces. As advances in technology continue to boost our ability to communicate in new and different ways, military forces are restructuring to incorporate these technologies. Yet we are still limited in our ability to measure the contributions made by information networks. We describe three recent studies at the Naval Postgraduate School that involve information networks. First, we examine a simulation model expanded from a two-person, zero-sum game to explore how information superiority contributes to battlefield results and how sensitive it is to information quality. Second, we examine how network-enabled communications affect the logistics operations in a centralized receiving and shipping point. The results are intended to provide operational insights for terminal node operations within a sustainment base. Third, we explore how social networks might be incorporated into agent-based models representing civilian populations in stability operations.

Revenue Management: Models and Methods
Kalyan T Talluri (ICREA and Universitat Pompeu Fabra), Garrett J van Ryzin (Columbia University), Itir Z Karaesmen (University of Maryland) and Gustavo J Vulcano (New York University)

Abstract:
Revenue management is the collection of strategies and tactics firms use to scientifically manage demand for their products and services. The practice has grown from its origins in airlines to its status today as a mainstream business practice in a wide range of industry areas, including hospitality, energy, fashion retail, and manufacturing. This article provides an introduction to this increasingly important subfield of operations research, with an emphasis on use of simulation. Some of the contents are based on excerpts from the book The Theory and Practice of Revenue Management (Talluri and van Ryzin 2004a), written by the first two authors of this article.

Verification and Validation of Simulation Models
Robert G. Sargent (Syracuse University)

Abstract:
In this paper we discuss verification and validation of simulation models. Four different approaches to deciding model validity are described; two different paradigms that relate verification and validation to the model development process are presented; various validation techniques are defined; conceptual model validity, model verification, operational validity, and data validity are discussed; a way to document results is given; a recommended procedure for model validation is presented; and model accreditation is briefly discussed.

Approximate Zero-Variance Simulation
Pierre L'Ecuyer (DIRO, Université de Montréal) and Bruno Tuffin (IRISA / INRIA)

Abstract:
Monte Carlo simulation applies to a wide range of estimation problems, but converges rather slowly in general. Variance reduction techniques can lower the estimation error, sometimes by a large factor, but rarely change the convergence rate of the estimation error. This error usually decreases as the inverse square root of the computational effort, as dictated by the central limit theorem. In theory, there exist simulation estimators with zero variance, i.e., that always provide the exact value. The catch is that these estimators are usually much too difficult (or virtually impossible) to implement. However, there are situations, especially in the context of rare-event simulation, where the zero-variance simulation can be approximated well enough to provide huge efficiency gains. Adaptive versions can even yield a faster convergence rate, including exponential convergence in some cases. This paper gives a brief overview of these methods and discuss their practicality.

Inside Discrete-Event Simulation Software: How it Works and Why it Matters
Thomas J. Schriber (University of Michigan) and Daniel T. Brunner (Kiva Systems, Inc.)

Abstract:
This paper provides simulation practitioners and consum-ers with a grounding in how discrete-event simulation software works. Topics include discrete-event systems; entities, resources, control elements and operations; simulation runs; entity states; entity lists; and entity-list management. The implementation of these generic ideas in AutoMod, SLX, and Extend is described. The paper con-cludes with several examples of “why it matters” for modelers to know how their simulation software works, including coverage not only of AutoMod, SLX, and Extend, but also of SIMAN (Arena), ProModel, and GPSS/H.

Guidelines for Commercial Off-the-Shelf Simulation Package Interoperability
Simon J. E. Taylor (Brunel University), Stephen J. Turner (Nanyang Technological University) and Steffen Strassburger (Ilmenau University of Technology)

Abstract:
Commercial-off-the-shelf (COTS) Simulation Packages (CSPs) are widely used visual interactive modeling environments such as Arena™, Anylogic™, Flexsim™, Simul8™, Witness™, etc. CSP Interoperability (or distributed simulation) is a technique that allows a simulation to be executed over several computers or for several simulations running on different computers to run together. This also relates to simulation languages such as SLX™ and GPSS/H™. There have been various attempts to interoperate these CSPs, some with the IEEE 1516 High Level Architecture (HLA). These can be quite complex and it is easy to loose track of exactly what is occurring between interoperating CSPs and their models. This paper introduces a set of Interoperability Reference Models (IRMs), or design patters for CSP Interoperability, that can be used as guidelines to simplify the interoperability process.

Approximate Dynamic Programming: Lessons From the Field
Warren B Powell (Princeton University)

Abstract:
Approximate dynamic programming is emerging as a powerful tool for certain classes of multistage stochastic, dynamic problems that arise in operations research. It has been applied to a wide range of problems spanning complex financial management problems, dynamic routing and scheduling, machine scheduling, energy management, health resource management, and very large-scale fleet management problems. It offers a modeling framework that is extremely flexible, making it possible to combine the strengths of simulation with the intelligence of optimization. Yet it remains a sometimes frustrating algorithmic strategy which requires considerable intuition into the structure of a problem. There are a number of algorithmic choices that have to be made in the design of a complete ADP algorithm. This tutorial describes the author's experiences with many of these choices in the course of solving a wide range of problems.