22 Chapter 1. Introduction Thesis Outline Explores Multi-Objective Evolutionary Algorithms for the data-driven inference of Fault Tree models, resulting in the FT-MOEA algorithm Chapter 1 Chapter 2 Published at IEEE Transactions on Dependable and Secure Computing Explores the use of Confusion Matrix-based metrics to guide the convergence of FT-MOEA. Aims to improve robustness. Chapter 4 Submitted to Formal Methods for Industrial Critical Systems (FMICS2024) Proposes a framework for maintenance policy optimization using multistate deterioration models and Deep Reinforcement Learning. Published at the Prognostics and Health Management Society (PHME24) Chapter 7 Introduces the research context and motivation, provides background, defines the research gaps and questions, offers a methodology overview, outlines the main research contributions, and lists the research outcomes. Includes the general discussion, placing the results in perspective and reflecting on their contributions and limitations. Chapter 8 Concludes the dissertation and enlist recommendations for future research. Chapter 9 Part I. Data-driven Inference of Fault Tree models Part II. Multi-state deterioration modelling Part III. Maintenance optimisation of multistate components Discussion, Conclusions & Recommendations Explores the use of symmetries when available in the failure data-set. Proposes the toolchain SymLearn with FT-MOEA in the back-end. Chapter 3 Published at Computer Safety, Reliability, and Security (SAFECOMP2022) Models deterioration of sewer pipes accounting for severity levels. This model is based on discrete-time Markov chains and it is validated on a case study. Compares homogeneous and inhomogeneous time Markov chains and assess suitability for deterioration modelling in sewer pipe networks. Chapter 5 Published at European Safety and Reliability Conference (ESREL2022) Published at European Safety and Reliability Conference (ESREL2024) Chapter 6 Figure 1.13: Thesis outline: Overview.
RkJQdWJsaXNoZXIy MjY0ODMw