Multi-agent Deep Reinforcement Learning for Zero Energy Communities

Prasad, Amit; Dusparić, Ivana

doi:10.1109/isgteurope.2019.8905628

Cited by 40 publications

(18 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 4 summarizes the features and limitations of approaches to these applications. Prasad and Dusparic [92] introduced a MADRL model to deal with energy sharing problem in a zeroenergy community that comprises a collection of zero energy buildings, which have the total energy use over a year smaller than or equal to the renewables generation within each building. A deep RL agent is used to characterize each building to learn appropriate actions in sharing energy with other buildings.…”

Section: Madrl Applicationsmentioning

confidence: 99%

Deep Reinforcement Learning for Multiagent Systems: A Review of Challenges, Solutions, and Applications

Nguyen

Nahavandi

2020

IEEE Trans. Cybern.

804

309

View full text Add to dashboard Cite

Reinforcement learning (RL) algorithms have been around for decades and employed to solve various sequential decision-making problems. These algorithms however have faced great challenges when dealing with high-dimensional environments. The recent development of deep learning has enabled RL methods to drive optimal policies for sophisticated and capable agents, which can perform efficiently in these challenging environments. This paper addresses an important aspect of deep RL related to situations that require multiple agents to communicate and cooperate to solve complex tasks. A survey of different approaches to problems related to multi-agent deep RL (MADRL) is presented, including non-stationarity, partial observability, continuous state and action spaces, multi-agent training schemes, multi-agent transfer learning. The merits and demerits of the reviewed methods will be analyzed and discussed, with their corresponding applications explored. It is envisaged that this review provides insights about various MADRL methods and can lead to future development of more robust and highly useful multi-agent learning methods for solving real-world problems.

show abstract

Section: Madrl Applicationsmentioning

confidence: 99%

Deep Reinforcement Learning for Multiagent Systems: A Review of Challenges, Solutions, and Applications

Nguyen

Nahavandi

2020

IEEE Trans. Cybern.

804

309

View full text Add to dashboard Cite

show abstract

“…Battery/thermal storage control: Regarding renewable energy sharing among different buildings, Prasad and Dusparic developed a Deep Reinforcement Learning (i.e., a machine learning approach that enables intelligent agents to learn the optimal behavior via trial and error)-based method for the ZEB community, with the purpose of reducing energy losses due to transmission and storage, and achieving economic gains [114]. Fan et al, also developed a collaborative DR control of zero energy buildings for building group performance improvements, in which the control of each building was conducted in sequence, and the optimization of one building's operation was based on the previously optimized buildings' operation, i.e., the optimization of (k+1)th building's operation is based on the aggregated operation of the first to kth buildings [113].…”

Section: Coordinated Controlsmentioning

confidence: 99%

A Technical Review of Modeling Techniques for Urban Solar Mobility: Solar to Buildings, Vehicles, and Storage (S2BVS)

et al. 2020

View full text Add to dashboard Cite

The deployment of solar photovoltaics (PV) and electric vehicles (EVs) is continuously increasing during urban energy transition. With the increasing deployment of energy storage, the development of the energy sharing concept and the associated advanced controls, the conventional solar mobility model (i.e., solar-to-vehicles (S2V), using solar energy in a different location) and context are becoming less compatible and limited for future scenarios. For instance, energy sharing within a building cluster enables buildings to share surplus PV power generation with other buildings of insufficient PV power generation, thereby improving the overall PV power utilization and reducing the grid power dependence. However, such energy sharing techniques are not considered in the conventional solar mobility models, which limits the potential for performance improvements. Therefore, this study conducts a systematic review of solar mobility-related studies as well as the newly developed energy concepts and techniques. Based on the review, this study extends the conventional solar mobility scope from S2V to solar-to-buildings, vehicles and storage (S2BVS). A detailed modeling of each sub-system in the S2BVS model and related advanced controls are presented, and the research gaps that need future investigation for promoting solar mobility are identified. The aim is to provide an up-to-date review of the existing studies related to solar mobility to decision makers, so as to help enhance solar power utilization, reduce buildings’ and EVs’ dependence and impacts on the power grid, as well as carbon emissions.

show abstract

“…The authors report an energy cost saving of 5%. A recent 2018 paper by Prasad and Dusparic implement a multi-agent deep RL approach to enable homes to share energy with one another in order to minimize cost for the community as a whole [96]. The authors implemented a multi-agent DQN approach where each agent controlled a house.…”

Section: Smart Homes and The Electrical Gridmentioning

confidence: 99%

A review of reinforcement learning for autonomous building energy management

Mason

Grijalva

2019

Computers & Electrical Engineering

228

View full text Add to dashboard Cite

The area of building energy management has received a significant amount of interest in recent years. This area is concerned with combining advancements in sensor technologies, communications and advanced control algorithms to optimize energy utilization. Reinforcement learning is one of the most prominent machine learning algorithms used for control problems and has had many successful applications in the area of building energy management. This research gives a comprehensive review of the literature relating to the application of reinforcement learning to developing autonomous building energy management systems. The main direction for future research and challenges in reinforcement learning are also outlined.

show abstract

Multi-agent Deep Reinforcement Learning for Zero Energy Communities

Cited by 40 publications

References 14 publications

Deep Reinforcement Learning for Multiagent Systems: A Review of Challenges, Solutions, and Applications

Deep Reinforcement Learning for Multiagent Systems: A Review of Challenges, Solutions, and Applications

A Technical Review of Modeling Techniques for Urban Solar Mobility: Solar to Buildings, Vehicles, and Storage (S2BVS)

A review of reinforcement learning for autonomous building energy management

Contact Info

Product

Resources

About