1st Edition

Large-Scale Simulation Models, Algorithms, and Applications

By Dan Chen, Lizhe Wang, Jingying Chen Copyright 2012
    260 Pages 107 B/W Illustrations
    by CRC Press

    260 Pages 107 B/W Illustrations
    by CRC Press

    Large-Scale Simulation: Models, Algorithms, and Applications gives you firsthand insight on the latest advances in large-scale simulation techniques. Most of the research results are drawn from the authors’ papers in top-tier, peer-reviewed, scientific conference proceedings and journals.

    The first part of the book presents the fundamentals of large-scale simulation, including high-level architecture and runtime infrastructure. The second part covers middleware and software architecture for large-scale simulations, such as decoupled federate architecture, fault tolerant mechanisms, grid-enabled simulation, and federation communities. In the third part, the authors explore mechanisms—such as simulation cloning methods and algorithms—that support quick evaluation of alternative scenarios. The final part describes how distributed computing technologies and many-core architecture are used to study social phenomena.

    Reflecting the latest research in the field, this book guides you in using and further researching advanced models and algorithms for large-scale distributed simulation. These simulation tools will help you gain insight into large-scale systems across many disciplines.

    FUNDAMENTALS
    Introduction
    Background
    Organization of the Book

    Background and Fundamentals
    High Level Architecture and Runtime Infrastructure
    Cloning and Replication
    Simulation Cloning
    Summary of Cloning and Replication Techniques
    Fault Tolerance
    Time Management Mechanisms for Federation Community

    MIDDLEWARE AND SOFTWARE ARCHITECTURES
    A Decoupled Federate Architecture
    Problem Statement
    Virtual Federate and Physical Federate
    Inside the Decoupled Architecture
    Federate Cloning Procedure
    Benchmark Experiments and Results
    Summary
    Exploiting the Decoupled Federate Architecture

    Fault-Tolerant HLA-Based Distributed Simulations
    Introduction
    Decoupled Federate Architecture
    A Framework for Supporting Robust HLA-Based Simulations
    Experiments and Results
    Summary

    Synchronization in Federation Community Networks
    Introduction
    HLA Federation Communities
    Time Management in Federation Communities
    Synchronization Algorithms for Federation Community Networks
    Experiments and Results
    Summary

    EVALUATION OF ALTERNATIVE SCENARIOS
    Theory and Issues in Distributed Simulation Cloning
    Decision Points
    Active and Passive Cloning of Federates
    Entire versus Incremental Cloning
    Scenario Tree
    Summary

    Alternative Solutions for Cloning in HLA-Based Distributed Simulation
    Single-Federation Solution versus Multiple-Federation Solution
    DDM versus Non-DDM in Single-Federation Solution
    Middleware Approach
    Benchmark Experiments and Results
    Summary

    Managing Scenarios
    Problem Statement
    Recursive Region Division Solution
    Point Region Solution
    Summary

    Algorithms for Distributed Simulation Cloning
    Overview of Simulation Cloning Infrastructure
    Passive Simulation Cloning
    Mapping Entities
    Incremental Distributed Simulation Cloning
    Summary

    Experiments and Results of Simulation Cloning Algorithms
    An Application Example
    Configuration of Experiments
    Correctness of Distributed Simulation Cloning
    Efficiency of Distributed Simulation Cloning
    Scalability of Distributed Simulation Cloning
    Optimizing the Cloning Procedure
    Summary of Experiments and Results
    Achievements in Simulation Cloning

    APPLICATIONS
    Hybrid Modeling and Simulation of a Huge Crowd over an HGA
    Introduction
    Crowd Modeling and Simulation
    The Hierarchical Grid Architecture for Large Hybrid Simulation
    Hybrid Modeling and Simulation of Huge Crowd: A Case Study
    Experiments and Results
    Summary

    Massively Parallel M&S of a Large Crowd with GPGPU
    Introduction
    Background and Notation
    The Hybrid Behavior Model
    A Case Study of Confrontation Operation Simulation
    Confrontation Operation Simulation Aided by GP-GPU
    Summary

    Index

    Biography

    Dan Chen is a professor and director of the Scientific Computing Lab at the China University of Geosciences. His research interests include computer-based modeling and simulation, high performance computing, and neuroinformatics.

    Lizhe Wang is a professor at the Center for Earth Observation and Digital Earth, Chinese Academy of Sciences. Dr. Wang is also a "ChuTian Scholar" Chair Professor at the China University of Geosciences, a senior member of IEEE, and a member of ACM. His research interests include high performance computing, grid/cloud computing, and data-intensive computing.

    Jingying Chen is a professor in the National Engineering Centre for e-Learning at Huazhong Normal University. Her research interests include intelligent systems, computer vision, and pattern recognition.