TY - GEN
T1 - Reinforcement learning for the unit commitment problem
AU - Dalal, Gal
AU - Mannor, Shie
N1 - Publisher Copyright: © 2015 IEEE.
PY - 2015/8/31
Y1 - 2015/8/31
N2 - In this work we solve the day-ahead unit commitment (UC) problem, by formulating it as a Markov decision process (MDP) and finding a low-cost policy for generation scheduling. We present two reinforcement learning algorithms, and devise a third one. We compare our results to previous work that uses simulated annealing (SA), and show a 27% improvement in operation costs, with running time of 2.5 minutes (compared to 2.5 hours of existing state-of-the-art).
AB - In this work we solve the day-ahead unit commitment (UC) problem, by formulating it as a Markov decision process (MDP) and finding a low-cost policy for generation scheduling. We present two reinforcement learning algorithms, and devise a third one. We compare our results to previous work that uses simulated annealing (SA), and show a 27% improvement in operation costs, with running time of 2.5 minutes (compared to 2.5 hours of existing state-of-the-art).
KW - Learning (artificial intelligence)
KW - Optimal scheduling
KW - Optimization methods
KW - Power generation dispatch
UR - http://www.scopus.com/inward/record.url?scp=84951335636&partnerID=8YFLogxK
U2 - 10.1109/PTC.2015.7232646
DO - 10.1109/PTC.2015.7232646
M3 - منشور من مؤتمر
T3 - 2015 IEEE Eindhoven PowerTech, PowerTech 2015
BT - 2015 IEEE Eindhoven PowerTech, PowerTech 2015
T2 - IEEE Eindhoven PowerTech, PowerTech 2015
Y2 - 29 June 2015 through 2 July 2015
ER -