2003-2004 ANNUAL REPORT

The Auburn Modeling and Simulation Laboratory of the M&SNet

Computer Science & Engineering

College of Engineering

Auburn University

Auburn, AL 36849

The objective of the Modeling & Simulation Laboratory is to study and develop new techniques of applying computer simulation.  A specific area of expertise is in the modeling and simulation of computer network and agent-directed simulation. AMSL plays a key role in supporting the activities of both Auburn's center for Innovations in Mobile, Pervasive, and Agile Computing Technologies (IMPACT) and the College of Engineering's wireless curriculum.

Specific areas of concentration include:

Simulation of computer networks, application of NS2 and OPNET to networking problems,  agent-directed simulation, simulation theory and methodology for conflict studies, model and simulation composability, combat simulation, interoperability of networked systems, and software architectural design for interoperability.

RECENT PUBLICATIONS BY AMSL MEMBERS

Technical/Final Reports and Other Presentations

RESEARCH ACTIVITIES and PROJECTS

Agent-Mediated Active Model Marketplaces on the Semantic Grid: Increased use of simulation modeling along with continuous evolution of model bases are expected to impose a burden in effective model discovery and reuse. Two emergent issues are (1) the lack of dynamic brokering mechanisms among developers and constantly evolving set of services provided by model providers and (2) the lack of high precision matchmaking methods over explicitly defined contextual model assumptions. It is widely accepted that the complexity of purpose and function involves the context within which a simulation is developed, composed, and used. Since implicit assumptions significantly complicate model reuse, existing repositories need to be augmented by mechanisms that operate on models’ conceptual, realization, and experimental contexts. To this end, the notion of an active model marketplace is introduced. The architecture promotes having model producers and consumers represented by agents providing services to one another under various forms of contracts in agent-mediated model marketplaces. A service-oriented agent-based marketplace strategy is being developed to facilitate development of model marketplaces that can permit large number of model providers and consumers, operating on rapidly evolving model bases, to share models more effectively than by using conventional repositories.

Dynamic Model and Simulation Updating in Multisimulation: In most realistic complex problems and conflicts, the nature of the problem changes as the simulation unfolds. Initial parameters, as well as models can be irrelevant under emergent conditions. Relevant contingency models need to be identified and instantiated to continue exploration. Furthermore, our knowledge about the problem (i.e., military conflict) being studied may not be captured by any single model or experiment. Novel multimodel and multisimulation methodologies are being explored to allow simulation of several aspects of reality simultaneously, to use coupled or atomic multimodels and/or regular models, to include simulation update in its basic form, and to have the branching possibility to start new simulation studies to be executed concurrently with at least one other alternative simulation study to experiment with several aspects of reality.

Integrating Model-based Verification into Software design Education: Proper design analysis and verification is indispensable to assure quality and reduce emergent costs due to faulty software. Despite the importance of verification, skills involved in critical analysis are often elusive in software engineering undergraduate curricula. This project focuses on certain pervading issues at the educational front to suggest a strategy that facilitates improvement of analysis and design evaluation skills within the context of software design education. The premise of the strategy is based on the observation that fundamental component of any engineering curriculum is the inclusion of formal and sound techniques that facilitate analysis of the artifacts produced by students. In software engineering curriculum, though, the lack of tool support and training in the use of formal or, at least, semi-formal methods is a significant barrier in teaching formal analysis. As such, we are exploring how selective and pragmatic application of formal methods can be integrated with undergraduate software design education. Tool support for the strategy is being developed to bridge the gap between the mathematical underpinnings of formal methods and student’s semi-formal design worldview.

Tools for Testing and Certification of Distributed, Dynamic Configurations of a Total Ship Computing Environment: Complex, dynamically reconfigurable, distributed systems are common in today’s component-based mission-critical systems. Conventional testing tools used for certifying such systems are less likely to succeed in such less-predictable and evolving domains where underlying assumptions with respect to the structure of the system are constantly being violated. Furthermore, given the vast complexity of such systems and impracticality of exhaustive testing of the overall multidimensional state space, new methodologies and computational infrastructures are necessary to deal with dynamically reconfigurable systems. We are developing a computational framework that could facilitate scalable and sound reasoning approach by localizing the detection of faults that occur at run-time. The premises of the approach are (1) local certification of components with respect to their contractual specifications, (2) packaging of the contracts, test cases, and built-in self test mechanisms with the component using a metadata wrapper technology, and (3) use of associated run-time interface violation detector mechanisms to detect deviations from acceptable behavior as components reconfigure and interact at run-time. By embodying the behavioral interface models with semi-automatically generated wrappers, the technology will support services for run-time contract violation checking and built-in-self testing even when the underlying component technology does not provide built-in introspective reflection capabilities.

CURRENT and PENDING AMSL GRANTS

Contracts/Grants

Preproposal Submittals

ACTIVITIES ORGANIZED BY AMSL

ANNOUNCEMENT: COOPERATION OPPORTUNITIES IN

CONFLICT SYSTEMS SIMULATION

A recent research effort on simulation modeling of conflict systems is initiated.

Topics*: Conflicts are social phenomena that are worth studying because they affect quality of life everywhere. New advanced simulation techniques may offer proper means to model and explore alternative and unforeseen consequences of conflicts. Political, economic, military as well as terrorist conflicts are the most destructive elements of the modern world. Proper methodologies are needed to conceive realistic models of complex conflict systems, the behavioral trajectory of which is never fixed due to uncertainty in various phases of the problem.

Positions: Auburn Modeling and Simulation Laboratory is seeking graduate students (M.S. and/or Ph.D. level) and visitors within M&SNet organizations to collaborate on a wide range of methodological, theoretical, and applied simulation modeling problems regarding conflict analysis, resolution, and management. We are looking for students, as well as visitors, who can contribute to our well-funded research program. The durations for the positions are: 2 years for M.S., 3-4 years for Ph.D., and up to 1 year for visitors.Applicants with interest and knowledge in agent-directed simulation and applications of simulation modeling to social science problems as well as to human behavior and conflicts are encouraged to contact Dr. Levent Yilmaz by email at:

yilmaz@eng.auburn.edu

* On these topics, AMSL is already cooperating with Dr. Tuncer Ören of the OC-MISS of M&SNet