Ingeniería Civil Matemática certificada por 3 años hasta Diciembre de 2026
Cristóbal Alexis Reyes Pavez
Bachelor’s Thesis
Reinforcement Learning for Multi-Component System Repair Policies
Thesis Advisor
Guido Lagos
Héctor Olivero
Sumary
The reliability of a system is defined as its ability to fulfill its purpose, operating continuously without interruptions. To ensure this stability, preventive maintenance strategies are implemented to maximize uptime and minimize both unexpected failures and the costs associated with repairs.
This work focuses on the optimization of repair policies in stochastic environments, particularly in systems with multiple components with random lifetimes. To address the problem, various lifetime distributions will be analyzed, and Markov decision processes will be used as the theoretical framework for optimization.
The effectiveness of the proposed strategies will be evaluated through simulations, demonstrating their impact on improving maintenance management. Using machine learning algorithms, strategies are designed to optimize system availability and reduce maintenance-related costs.
