Using multi-agent Q learning to select best possible action for each of non-shiftable, power-shiftable and time-shiftable appliances.

Demonstration of cost saving: main.ipynb

References:

• Xu, Xu & Jia, Youwei & Xu, Yan & Xu, Zhao & Chai, Songjian & Lai, Chun Sing. (2020). A Multi-Agent Reinforcement Learning-Based Data-Driven Method for Home Energy Management. IEEE Transactions on Smart Grid. PP. 1-1. 10.1109/TSG.2020.2971427.
• Jiménez-Bravo, Diego & Pérez-Marcos, Javier & Hernández de la Iglesia, Daniel & Villarrubia, G. & Paz, Juan. (2019). Multi-Agent Recommendation System for Electrical Energy Optimization and Cost Saving in Smart Homes. Energies. 12. 1317. 10.3390/en12071317.

Using exploring actions according to Q-value ratios (EAQR) to coordinate power consumption between appliances while considering a power limit.

Working of EAQR: eaqr.ipynb

Example showing use of power limit: power-limit.ipynb

Reference:

• Wang, Wenbo & Zhang, Zhen & Wang, Dongqing. (2018). EAQR: A Multiagent Q-Learning Algorithm for Coordination of Multiple Agents, 10.1155/2018/7172614, Complexity