hpi-sam / rl-4-self-repair

Reinforcement Learning Models for Online Learning of Self-Repair and Self-Optimization
MIT License
0 stars 1 forks source link

Design Failure fix outcome based on component dependency data #24

Open christianadriano opened 4 years ago

christianadriano commented 4 years ago

Currently, the failure to fix a component is given by a static parameter. We will change to compute the failure to fix based on the dependencies that a component has on other components that might be failing.

We can represent the dependencies and the probabilities using a table, where: First column corresponds to the component that we want to fix Other columns correpond to the potential dependencies to other components.

The cells of the table will provide the probability of failing to fix (from 0% to 100%) if a given component is also failing. See excel attached. MarkovChain_FailingToFix_Transitions.xlsx