[Paper Review] Peer-to-Peer Energy Sharing: A New Business Model Towards a Low-Carbon Future

Yue Chen, Changhong Zhao|arXiv (Cornell University)|Aug 5, 2021

Smart Grid Energy Management100 references4 citations

TL;DR

This paper proposes peer-to-peer (P2P) energy sharing as a transformative business model to enable prosumers—households generating and consuming energy—to trade electricity directly, enhancing grid efficiency and supporting a low-carbon future. By leveraging decentralized market mechanisms, smart contracts, and privacy-preserving algorithms, the model optimizes energy exchange, reduces reliance on centralized storage, and achieves near-social-optimal outcomes while preserving participant privacy and addressing conflicting incentives.

ABSTRACT

The development of distributed generation technology is endowing consumers the ability to produce energy and transforming them into "prosumers". This transformation shall improve energy efficiency and pave the way to a low-carbon future. However, it also exerts critical challenges on system operations, such as the wasted backups for volatile renewable generation and the difficulty to predict behavior of prosumers with conflicting interests and privacy concerns. An emerging business model to tackle these challenges is peer-to-peer energy sharing, whose concepts, structures, applications, models, and designs are thoroughly reviewed in this paper, with an outlook of future research to better realize its potentials.

Motivation & Objective

• To analyze the structural, functional, and economic foundations of P2P energy sharing as a scalable solution for integrating distributed energy resources (DERs).
• To identify and categorize business models, market mechanisms, and design principles that enable efficient, privacy-preserving, and incentive-compatible energy trading among prosumers.
• To address challenges in system operation caused by volatile renewable generation, conflicting prosumer interests, and information asymmetry.
• To explore future research directions in multi-operator markets, multi-energy systems, and decision-dependent uncertainty in renewable energy forecasting.

Proposed method

• The paper conducts a comprehensive literature review of energy sharing, peer-to-peer trading, transactive energy, and sharing economy models, synthesizing findings across 20+ key studies.
• It classifies energy sharing systems into three structural types: centralized, distributed, and decentralized, each with distinct control and communication architectures.
• It evaluates market mechanisms using game-theoretic models (e.g., Nash equilibrium) and optimization-based approaches (e.g., ADMM, dual decomposition) to achieve system-wide efficiency.
• It examines privacy-preserving techniques such as consensus algorithms and information revelation mechanisms to ensure truthful reporting without exposing sensitive data.
• It models energy flows across multiple energy carriers (e.g., electricity, hydrogen) and evaluates integrated energy market designs.
• It applies learning-based methods such as deep reinforcement learning and Q-networks to enable adaptive, real-time energy sharing in dynamic environments.

Experimental results

Research questions

• RQ1 How can P2P energy sharing systems be structured to ensure efficiency, scalability, and privacy in decentralized energy trading among prosumers?
• RQ2 What market mechanisms can align individual incentives with social welfare while preventing strategic behavior and information asymmetry?
• RQ3 How can multi-energy systems (e.g., electricity, hydrogen) be integrated into a unified energy sharing framework to enhance system flexibility?
• RQ4 What role can credit-based or non-financial value systems play in sustaining long-term participation without relying on subsidies?
• RQ5 How can decision-dependent uncertainty—where system decisions affect renewable output—be modeled and optimized in energy sharing frameworks?

Key findings

• P2P energy sharing significantly reduces reliance on backup energy storage by enabling direct, real-time exchange between prosumers, thereby minimizing waste from volatile renewable sources.
• Decentralized market mechanisms based on dual decomposition and ADMM converge to socially optimal equilibria while preserving participant privacy through iterative, distributed computation.
• Privacy-preserving consensus algorithms enable bidirectional energy exchange between electric vehicles and the grid without exposing sensitive load or generation data.
• Learning-based methods such as deep Q-networks and multi-agent reinforcement learning achieve high utility gains in microgrid and zero-energy community settings by adapting to dynamic prosumer behavior.
• Current models often assume symmetric information, but asymmetric information can lead to market failure unless mechanisms are designed to incentivize truthful reporting.
• Future systems must address decision-dependent uncertainty—where operational decisions (e.g., solar panel angle) influence renewable output—since this is currently underexplored and computationally challenging.