Distributed Simulation of Synthetic Environments and Wireless Networks
Abstract
The goal of this research is to provide a decision aid that allows engineers, scientists and battle managers to simulate complex situations and facilitate rapid and informed planning. Specifically, we are developing mechanisms to facilitate the decision process in battle management. Often, very large systems may need to be modeled such as an entire combat theater or a large airspace. The problem with large systems is that a single simulation run may take hours to complete. In order to reduce execution time, one can decompose a sequential simulation into a collection of many interacting programs distributed over multiple processors. This approach is called parallel simulation. To ensure correctness, we require that parallel execution generates the same result as a sequential execution. This technology enables off-line decision aid simulations to be transformed into real time decision aids for time critical situations without loss of model accuracy. To build fast, efficient and powerful simulators we need parallel algorithms that: (1) hide computational latencies for image generation in virtual environments, (2) allow effective what if and alternative scenario analysis and decision making in time critical situations, (3) maximize forward progress via background execution, (4) implement interactive visualization techniques that provide insight into parallel simulation execution to prevent potential bottlenecks, and (5) provide for geographically distributed (WAN) environments and exploit the capabilities of high performance interconnections to deliver real time results. To address these issues the development and integration of new mechanisms for interactive parallel discrete event simulations are presented, implemented and evaluated.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 03, 1999
- Accession Number
- ADA369488
Entities
People
- Richard M. Fujimoto
Organizations
- Georgia Tech