Maintenance Scheduling

Last Week

GUI Simulator testing
Final simulator idea -> Gantt Chart
Single agent + objective function (reward function)

GUI Maintenance Scheduling

Single agent vs Multi agent

Explosive Action Space

Su, J., Huang, J., Adams, S., Chang, Q., & Beling, P. A. (2022). Deep multi-agent reinforcement learning for multi-level preventive maintenance in manufacturing systems. Expert Systems with Applications, 192, 116323.
Concurrency





Planning

Masking actions
Remove the negative reward of invalid actions
Improve the training
Time steps to days
Technicians with capacity of hours per day
1 week scheduling
CEBI maintenance data

CEBI Data