Dynamic Data-Driven Simulation. Real-Time Data for Dynamic System Analysis and Prediction

✍ Scribed by Xiaolin Hu

• Publisher: World Scientific Publishing
• Year: 2023
• Tongue: English
• Leaves: 328
• Category: Library

✦ Table of Contents

Contents
Preface
About the Author
List of Tables
List of Figures
1. Introduction
1.1 Why This Book?
1.2 Scope and Structure of the Book
Part 1: Foundation
2. Dynamic System, Simulation, and Data
2.1 Dynamic System and Simulation
2.2 Framework for Modeling and Simulation
2.3 Offline Simulation and Online Simulation
2.4 The Rise of Data
2.5 Connecting Data to Simulation
2.5.1 Initial state
2.5.2 Model parameters
2.5.2.1 Model calibration and validation
2.5.3 External input
2.5.4 Simulation output
2.6 Combining Simulation Model and Real-Time Data
2.7 Data Modeling vs. Simulation Modeling
2.7.1 What is needed for developing a model? Data vs. system knowledge
2.7.2 What does a model represent? Correlation vs. causality
2.7.3 What can a model do? Predictive analysis vs. prescriptive analysis
2.8 Sources
3. Simulation Models and Algorithms
3.1 A Taxonomy of Simulation Models
3.1.1 Timing of change
3.1.2 Aggregation level
3.1.3 Randomness
3.2 A Structural View of Simulation Model and Simulator
3.3 Continuous Model and Simulation Algorithm
3.3.1 Continuous simulation example
3.4 Discrete Time Model and Simulation Algorithm
3.4.1 Discrete time simulation example
3.5 Discrete Event Model and Simulation Algorithm
3.5.1 Discrete Event System Specification (DEVS) model
3.5.2 Discrete event simulation algorithm
3.5.3 Discrete event simulation example
3.6 Agent-Based Model and Simulation Algorithm
3.6.1 Agent-based model
3.6.2 Agent-based simulation algorithm
3.6.3 Agent-based simulation example
3.7 Cellular Space Model
3.7.1 Cellular space simulation example
3.8 System Dynamics Model
3.8.1 System dynamics model example
3.9 Sources
4. Basic Probability Concepts
4.1 Random Variable and Probability Distribution
4.2 Joint Distribution, Marginal Distribution, and Conditional Probability
4.3 Expectation, Variance, and Covariance
4.4 Multivariate Distribution
4.5 Monte Carlo Method
4.6 Sources
Part 2: Dynamic Data-Driven Simulation
5. Dynamic Data-Driven Simulation
5.1 What is Dynamic Data-Driven Simulation?
5.2 DDDS Activities
5.3 Real-Time Data
5.4 Dynamic State Estimation
5.4.1 The state estimation problem
5.4.2 Data assimilation for state estimation
5.5 Online Model Calibration
5.6 External Input Modeling & Forecasting
5.6.1 Modeling with probability distribution
5.6.2 Simulation modeling
5.6.3 Time series analysis
5.6.3.1 Simple moving average
5.6.3.2 Simple exponential smoothing
5.7 Simulation-Based Prediction/Analysis
5.8 Relation to Other Modeling and Simulation Activities
5.9 Sources
6. Data Assimilation
6.1 Introduction
6.2 State Space and Belief Distribution
6.3 Probabilistic State Representation
6.3.1 Gaussian representation
6.3.2 Sample-based representation
6.3.3 Updating belief distribution from state transition
6.3.3.1 The sampling approach
6.3.3.2 The analytical approach
6.4 The Data Assimilation Problem
6.4.1 State-space formulation
6.4.1.1 The Markov assumption
6.4.1.2 Simulation model for the state-space formulation
6.4.1.3 Process noise and measurement noise
6.4.2 Filtering and smoothing
6.4.3 Joint state–parameter estimation
6.5 Sequential Bayesian Filtering
6.5.1 Mathematical derivation
6.5.2 The prediction–update procedure
6.5.3 The sequential Bayesian filtering algorithm
6.5.4 Illustrative example
6.6 Likelihood Computation from Measurement Data
6.6.1 Scalar Gaussian measurement
6.6.1.1 Univariate scalar Gaussian measurement
6.6.1.2 Multivariate scalar Gaussian measurement
6.6.2 Non-scalar Gaussian measurement
6.7 Kalman Filter
6.7.1 Linear Gaussian state-space model
6.7.2 The Kalman filter algorithm
6.7.3 Extensions of the Kalman filter
6.7.3.1 Extended Kalman filter
6.7.3.2 Ensemble Kalman filter
6.7.4 Illustrative example
6.8 Particle Filters
6.8.1 Importance sampling
6.8.2 Resampling
6.8.3 The particle filter algorithm
6.8.4 Illustrative example
6.9 Sources
7. Dynamic Data-Driven Simulation for Discrete Simulations
7.1 Introduction
7.2 A Dynamic Data-Driven Simulation Framework
7.3 Particle Filter-Based Data Assimilation
7.3.1 Sampling using discrete simulation models
7.3.2 Measurement data and measurement model
7.3.3 Particle rejuvenation
7.4 Identical Twin Experiment
7.5 A Tutorial Example
7.5.1 The one-way traffic control system
7.5.2 The discrete event simulation model
7.5.3 Experiment design
7.5.4 Particle-filter-based data assimilation
7.5.5 Dynamic state estimation
7.5.6 Online model calibration
7.5.7 Simulation-based prediction/analysis
7.5.8 Data assimilation performance
7.6 Sources
Part 3: Application and Look Ahead
8. The Wildfire Spread Simulation Application
8.1 Introduction
8.2 DDDS for Wildfire Spread Simulation
8.2.1 Data assimilation
8.2.2 Wildfire spread simulation model
8.2.3 Real-time wildfire measurement data
8.2.4 External inputs modeling and forecasting
8.2.5 Simulation-based prediction/analysis
8.3 The DEVS-FIRE Simulation Model
8.4 Temperature Measurement Data
8.5 Data Assimilation for Dynamic State Estimation
8.5.1 Problem formulation
8.5.2 Sampling using the state transition model
8.5.3 Measurement model
8.5.4 The particle-filter-based data assimilation algorithm
8.5.5 Data assimilation results
8.6 Spatial Partition-Based Particle Filter
8.6.1 Motivation
8.6.2 The spatial partition-based particle filter
8.6.3 Experiment results
8.7 Parallel/Distributed Particle Filtering
8.8 Sources
9. Look Ahead
9.1 A New Area of Research and Development
9.2 Simulation Model
9.3 Measurement Data
9.4 Data Assimilation
9.5 External Inputs
9.6 Computation Cost
9.7 DDDS Applications
References
Index