Multi-Agent Deep Reinforcement Learning for Multi O... | AI Research

Key Takeaways

Multi-Agent Deep Reinforcement Learning for Multi Objective Battery Management in Dairy Farms The dairy industry in Ireland holds significant potential for r...
The dairy industry in Ireland has a large potential for the integration of renewable energy and the reduction of carbon emissions.
However, researchers of distributed generation control are mainly focused on residential and commercial applications.
The proposed control is organised in two layers: the upper layer uses dynamic pricing, and the lower layer is based on multi-agent reinforcement learning for battery management.
This paper also simulates the electrical response of the proposed control system in a rural distribution circuit.

Paper AbstractExpand

The dairy industry in Ireland has a large potential for the integration of renewable energy and the reduction of carbon emissions. However, researchers of distributed generation control are mainly focused on residential and commercial applications. To contribute to the effective integration of renewable energy in the dairy sector, this paper presents a multi-objective optimisation control system based on differential evolution and multi agent Deep Reinforcement Learning. The proposed control is organised in two layers: the upper layer uses dynamic pricing, and the lower layer is based on multi-agent reinforcement learning for battery management. This paper also simulates the electrical response of the proposed control system in a rural distribution circuit. The simulation results show that the proposed control framework can improve profits from energy arbitrage up to 18% compared to using Rule-based models, increase the use of distributed generation without significantly increasing cost, and comply with the Irish grid code in terms of voltage variation.

Multi-Agent Deep Reinforcement Learning for Multi Objective Battery Management in Dairy Farms

The dairy industry in Ireland holds significant potential for reducing carbon emissions through the integration of renewable energy. While much of the existing research on distributed energy control focuses on residential or commercial settings, this paper addresses the specific needs of the dairy sector. The authors propose a multi-objective optimization control system designed to manage energy effectively, helping dairy farms transition toward greener operations while maintaining economic viability.

A Two-Layered Control Strategy

To manage energy effectively, the researchers developed a control framework organized into two distinct layers. The upper layer utilizes dynamic pricing to inform energy decisions, while the lower layer employs multi-agent Deep Reinforcement Learning (MARL) to manage battery storage. By combining this reinforcement learning approach with differential evolution, the system is able to balance multiple objectives, such as maximizing profit and optimizing the use of renewable energy sources.

Improving Energy Efficiency and Profitability

The researchers tested their framework by simulating its electrical response within a rural distribution circuit. The results demonstrate that this multi-agent approach outperforms traditional rule-based models. Specifically, the system achieved up to an 18% increase in profits from energy arbitrage. Furthermore, the framework successfully increased the utilization of distributed renewable energy without causing a significant rise in operational costs.

Compliance and Grid Stability

Beyond financial gains, the study highlights the importance of grid stability. A key requirement for any energy management system is adherence to local regulations. The simulation results confirm that the proposed control framework complies with the Irish grid code, particularly regarding voltage variation. This ensures that the integration of renewable energy and battery management on dairy farms does not negatively impact the reliability or quality of the local electrical distribution network.