Deep reinforcement learning driven inspection and maintenance planning under incomplete information and constraints

Keyphrases

• Incomplete Information 100%
• Learning-based 100%
• Maintenance Scheduling 100%
• Deep Reinforcement Learning (deep RL) 100%
• Multi-agent Deep Reinforcement Learning (MADRL) 100%
• Long-run Risk 50%
• Stochasticity 50%
• Curse of Dimensionality 50%
• Parametrization 50%
• Number of Components 50%
• Resource Base 50%
• Computational Challenges 50%
• System Response 50%
• Decision Tree 50%
• Budget Constraint 50%
• Bayesian Inference 50%
• Adept 50%
• Maintenance Policy 50%
• Long-term Costs 50%
• Exponential Scaling 50%
• Lagrangian Relaxation 50%
• Partially Observable Markov Decision Process 50%
• Resource Scarcity 50%
• Risky Driving 50%
• State Uncertainty 50%
• Engineering Environment 50%
• State Action 50%
• Inspection Policy 50%
• Stochastic Dynamic Programming 50%
• Decentralized Control 50%
• Complex Optimization Problems 50%
• Risk-aware 50%
• Joint Framework 50%
• Action Sets 50%
• Intervention Engagement 50%

Computer Science

• Incomplete Information 100%
• Deep Reinforcement Learning 100%
• Multi-Agent Deep Reinforcement Learning 100%
• Optimization Problem 50%
• Parametrization 50%
• Cardinality 50%
• System Response 50%
• Stochastic Dynamic Program 50%
• Decentralized Control 50%
• Markov Decision Process 50%
• Lagrangian Relaxation 50%
• Decision Tree 50%
• Algorithmic Step 50%