REINFORCE for Predictive Maintenance

Experiment with the simple naïve REINFORCE algorithm
Run and compare with industry gold-standard OpenAI's Stable-Baseline-3 implementations of DQN, A2C and PPO
4 main phases of development and experiments
- Simple environment with a single env. variable > the Tool wear (mm)
- A simulated Dasic (2006) milling tool wear model
- PHM 2010 prognostics data
- Complex environment with features from the PHM 2010 data-set
Metrics

Name		Name	Last commit message	Last commit date
Expt_Reports_R1R2R3		Expt_Reports_R1R2R3
data		data
storage		storage
.gitignore		.gitignore
Experiments_V8.csv		Experiments_V8.csv
PdM_REINFORCE_ModelTester_V3.py		PdM_REINFORCE_ModelTester_V3.py
PdM_REINFORCE_V.10.RR.ipynb		PdM_REINFORCE_V.10.RR.ipynb
PdM_REINFORCE_V4.6.py		PdM_REINFORCE_V4.6.py
README.md		README.md
REINFORCE_V1 (Original_LG).ipynb		REINFORCE_V1 (Original_LG).ipynb
REINFORCE_V2 (Revised_RS).ipynb		REINFORCE_V2 (Revised_RS).ipynb
TestModels.csv		TestModels.csv
Visualizations_RF_for_PdM.ipynb		Visualizations_RF_for_PdM.ipynb
arn.bat		arn.bat
atr.bat		atr.bat
concat.py		concat.py
milling_tool_environment.py		milling_tool_environment.py
reinforce_classes.py		reinforce_classes.py
utilities.py		utilities.py

Navigation Menu