Temporal Robustness in Discrete Time Linear Dynamical Systems
Nilava Metya, Arunesh Sinha·May 05, 2025
Summary
A robust cost estimation method for discrete time linear systems using Markov chains in a Wasserstein ambiguity set is presented. It ensures global asymptotic stability, addressing uncertainty in time horizons. Applications include disease spread modeling and cyber-security. The paper also explores Markov state models, Monte Carlo methods, and their applications in healthcare, system identification, and dynamic health policy.
Introduction
Background
Overview of discrete time linear systems
Importance of cost estimation in system analysis
Challenges in estimating costs under uncertainty
Objective
To present a robust cost estimation method using Markov chains in a Wasserstein ambiguity set
To ensure global asymptotic stability in the presence of uncertainty in time horizons
Method
Data Collection
Gathering data on system parameters and uncertainties
Data Preprocessing
Cleaning and formatting data for analysis
Markov State Models
Definition and properties of Markov state models
Monte Carlo Methods
Application of Monte Carlo simulations for uncertainty quantification
Integration of Markov Chains and Wasserstein Ambiguity Set
Methodology for incorporating Markov chains within a Wasserstein ambiguity set
Global Asymptotic Stability
Techniques for ensuring stability in the estimation process
Applications
Disease Spread Modeling
Case studies in epidemiology
Benefits and limitations of the proposed method
Cyber-Security
Application in network security
Analysis of system vulnerabilities and robustness
Healthcare
Use in healthcare system optimization
Impact on resource allocation and patient care
System Identification
Role in identifying system parameters
Comparison with traditional methods
Dynamic Health Policy
Application in policy-making
Decision support for healthcare management
Conclusion
Summary of Findings
Implications for Future Research
Recommendations for Practitioners
Basic info
papers
optimization and control
artificial intelligence
Advanced features