This repository contains scripts that use reinforcement
learning to determine the optimal policy for any given
state in a wind plus storage system. The input data
files come from the Electric Reliability Council of
Texas (ERCOT).These are formatted using
data_processing.py
and then output into
state_space.csv
. Then, plant.py
uses Q-learning to
find the optimal policy for each data point, and
outputs the optimal policy to policies.csv
. Lastly,
results_analysis.py
reads in the policies and compares
the performance of the policy to a baseline of a system
with no storage included.
underwood-scott / cs-238-final-project Goto Github PK
View Code? Open in Web Editor NEWFinal project for CS 238 - Decision Making Under Uncertainty at Stanford University