Vodenčarević, Asmir: Identifying behavior models for hybrid production systems. 2013
Content
- Introduction
- Part I Background
- Part II Complexity of Automata Identification
- Complexity of Identifying Deterministic Automata
- Introduction
- Three Classes of Deterministic Automata
- Automaton Identification Problem
- Identification in the Limit
- Polynomial Identification in the Limit
- Conclusion
- Complexity of Identifying Stochastic Deterministic Automata
- Introduction
- Notations and Automata Definitions
- Identification in the Limit with Probability One
- Strong Polynomial Criteria for Identification
- Weak Polynomial Criteria for Identification
- Summary
- Polynomial Approximations of Stochastic Automata
- Part III Algorithms
- Automated Learning of 1-SDHAs from Data
- Data Acquisition and Preprocessing
- Generating Alphabet and Timing Constraints from Measurements
- The HyBUTLA Learning Algorithm
- Abrupt Change Detection
- Modeling Autonomous Jumps with State Splits
- Algorithm Properties
- Conclusion
- Anomaly Detection Based on Learned Behavior Models
- Part IV Case Studies in Learning and Anomaly Detection
- Real-World Plants
- Comparative Empirical Analysis on Learning Automata
- Learning Behavior Models for the Lemgo Model Factory
- Anomaly Detection Experiments
- Conclusion
- Artificial Datasets
- Part V Conclusion