Explanation-Driven Self-Adaptations (XDA) Replication Package

Our papers

• Efficient Self-Adaptation through Explanation-Driven White-Box Optimization
• Explanation-driven Self-adaptation using Model-agnostic Interpretable Machine Learning

Install Requirements

pip install -r requirements.txt

For macOS and Linux Users

chmod +x MDP_Dataset_builder/evaluate_adaptations.sh

Generate Dataset (Optional)

Inside MDP_Dataset_builder/run.sh and MDP_Dataset_builder/run.bat:

• MAX_SAMPLES: number of samples to generate
• TOTAL_THREADS: number of threads to use for the generation

For macOS and Linux Users

chmod +x MDP_Dataset_builder/run.sh
./run.sh

For Windows Users

.\run.bat

Run Adaptation Tests

Inside main/main.py:

• line 43: you can specify the path to your dataset
• line 61: you can specify the list of requirements to consider
• line 64: you can specify the size of the neighborhood
• line 65: you can specify the number of starting solutions to consider
• line 68: you can specify the target success probabilities for each requirement
• line 120: you can specify the number of tests to do

python main/main.py

Generate Plots

python main/makeAllPlots.py