Data-Driven Learning and Load Ensemble Control

Ali Hassan, Deepjyoti Deka, Michael Chertkov, Yury Dvorkin

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)


Demand response (DR) programs aim to engage distributed small-scale flexible loads, such as thermostatically controllable loads (TCLs), to provide various grid support services. Linearly Solvable Markov Decision Process (LS-MDP), a variant of the traditional MDP, is used to model aggregated TCLs. Then, a model-free reinforcement learning technique called Z-learning is applied to learn the value function and derive the optimal policy for the DR aggregator to control TCLs. The learning process is robust against uncertainty that arises from estimating the passive dynamics of the aggregated TCLs. The efficiency of this data-driven learning is demonstrated through simulations on Heating, Cooling & Ventilation (HVAC) units in a testbed neighborhood of residential houses.

Original languageEnglish
Article number106780
JournalElectric Power Systems Research
Publication statusPublished - Dec 2020
Externally publishedYes


  • Linearly Solvable MDP
  • Markov Decision Process
  • TCL ensemble
  • Thermostatically Controlled Loads
  • Z-learning


Dive into the research topics of 'Data-Driven Learning and Load Ensemble Control'. Together they form a unique fingerprint.

Cite this