DRL policy for Bridge-maintenance

• License: MIT
• Copyright: Copyright (c) SMC http://smc.hit.edu.cn/
  Structural Monitoring & Control
  School of Civil Engineering
  Harbin Institute of Technology
  

About the cproject

• DRL policy establishes a Deep Reinforcement Learning (DRL) framework for structural maintenance management by reformulating the problem as a standard control problem.
• With the great power of nonlinear representation of the DNNs, it is able to handle the decision-making in structural maintenance of large complex structures.
• The code is the detail implementation of paper Optimal policy for structure maintenance: A deep reinforcement learning framework.
• Two cases are provided here, the deck-system (for simple structures), and the long-span cable-stayed bridge (for complex structures).