Ann S. Yoshimura

The IDES framework: a case study in development of a parallel discrete-event simulation system

This tutorial describes considerations in the design and development of the IDES parallel simulation system. IDES is a Java-based parallel/distributed simulation system designed to support the study of complex large-scale enterprise systems. Using the IDES system as an example, the authors discuss how anticipated model and system constraints molded the design decisions with respect to modeling, synchronization, and communication strategies.

Performance modeling of the IDES framework

The IDES project at Sandia National Laboratories is developing a large scale portable parallel simulator for use in stockpile stewardship. IDES will use the Breathing-Time-Buckets synchronization protocol; to support IDES development, this paper studies a performance model and describes performance experiments on expected work load and architectural parameters. A new parallel algorithm for terminating the window quickly is also described and analyzed.

The role of integrated modeling and simulation in disaster preparedness and emergency preparedness and response: The SUMMIT platform

The Department of Homeland Security, Science and Technology Directorate (DHS S&T), in partnership with the Federal Emergency Management Agency (FEMA), and in collaboration with Sandia National Laboratories, has developed the Standard Unified Modeling, Mapping, and Integration Toolkit (SUMMIT) as a geo-agile platform that enables users to discover and reuse models, integrate them quickly and economically, and apply them in analyses to improve the planning and execution of large and small exercises, planning efforts, and eventually response operations. The SUMMIT technology is innovative and promises to reduce the exercise cycle time by more than an order of magnitude. SUMMIT has been deployed to support National Level Exercises, regional exercises, and international exercises in Sweden, revolutionizing how exercises are designed and conducted. SUMMIT is now in the transition stage, from DHS S&T to the FEMA National Exercise and Simulation Center (NESC), where it will become available to the emergency management community. It will support the National Preparedness System[1] and integrate with the NESC Technology Platform in support of local, state, regional and national level exercises.

Using Performance Measurements to Improve MapReduce Algorithms

Abstract The Hadoop MapReduce software environment is used for parallel processing of distributively stored data. Data mining algorithms of increasing sophistication are being implemented in MapReduce, bringing new challenges for performance measurement and tuning. We focus on analyzing a job after completion, utilizing information collected from Hadoop logs and machine metrics. Our analysis, inspired by [1] [2] , goes beyond conventional Hadoop Job-Tracker analysis by integrating more data and providing web browser visualization tools. This paper describes examples where measurements helped diagnose subtle issues and improve algorithm performance. Examples demonstrate the value of correlating detailed information that is not usually examined in standard Hadoop performance displays.

Nonlinear optimization for stochastic simulations.

This report describes research targeting development of stochastic optimization algorithms and their application to mission-critical optimization problems in which uncertainty arises. The first section of this report covers the enhancement of the Trust Region Parallel Direct Search (TRPDS) algorithm to address stochastic responses and the incorporation of the algorithm into the OPT++ optimization library. The second section describes the Weapons of Mass Destruction Decision Analysis Center (WMD-DAC) suite of systems analysis tools and motivates the use of stochastic optimization techniques in such non-deterministic simulations. The third section details a batch programming interface designed to facilitate criteria-based or algorithm-driven execution of system-of-system simulations. The fourth section outlines the use of the enhanced OPT++ library and batch execution mechanism to perform systems analysis and technology trade-off studies in the WMD detection and response problem domain.