PaperlayerView exact revision on arXiv

Source-linked AI summary

Transforming Energy Networks via Peer to Peer Energy Trading: Potential of Game Theoretic Approaches

Wayes Tushar, Chau Yuen, Hamed Mohsenian-Rad, Tapan Saha, H. Vincent Poor, Kristin L Wood

arXiv:1804.00962v1eess.SPcs.GTeess.SY

TL;DR

P2P energy trading must model participants with conflicting interests and motivate participation and cooperation for smart-grid energy management. This paper systematically reviews game and auction theories, explains P2P trading and a testbed, and examines selected models. It concludes that these approaches have substantial potential for P2P energy trading while identifying future directions including game theory and blockchain-related privacy considerations.

  • Problem

    P2P trading requires decision-making models that account for participants’ conflicting interests and motivate participation and cooperation.

  • Method

    The paper systematically reviews game and auction approaches, introduces P2P trading features and a testbed, and discusses specific models used in P2P energy trading.

  • Results

    The paper reports important findings from selected game- and auction-theoretic P2P energy-trading studies.

  • Takeaways & Limitations

    Game-theoretic approaches have substantial potential for P2P energy trading and remain an important future research direction.

Abstract

from arXiv · show

Peer-to-peer (P2P) energy trading has emerged as a next-generation energy management mechanism for the smart grid that enables each prosumer of the network to participate in energy trading with one another and the grid. This poses a significant challenge in terms of modeling the decision-making process of each participant with conflicting interest and motivating prosumers to participate in energy trading and to cooperate, if necessary, for achieving different energy management goals. Therefore, such decision-making process needs to be built on solid mathematical and signal processing tools that can ensure an efficient operation of the smart grid. This paper provides an overview of the use of game theoretic approaches for P2P energy trading as a feasible and effective means of energy management. As such, we discuss various games and auction theoretic approaches by following a systematic classification to provide information on the importance of game theory for smart energy research. Then, the paper focuses on the P2P energy trading describing its key features and giving an introduction to an existing P2P testbed. Further, the paper zooms into the detail of some specific game and auction theoretic models that have recently been used in P2P energy trading and discusses some important finding of these schemes.

I. BACKGROUND AND MOTIVATION

P2P energy trading directly connects prosumers and the grid, but requires mechanisms that model conflicting decisions and motivate cooperation for efficient energy management. The paper surveys game- and auction-theoretic approaches as tools for addressing these challenges.

  • I. BACKGROUND AND MOTIVATION: FiT prosumers sell excess renewable energy to the grid and buy from it during shortages, but price differences limit their benefits.Some FiT techniques have consequently been discontinued.
  • I. BACKGROUND AND MOTIVATION: P2P trading allows small-scale participants to trade energy directly in real time, supporting community balance between generation and consumption.Consumers may buy renewable energy from peers with excess generation, reducing dependence on the grid or central supplier.
  • I. BACKGROUND AND MOTIVATION: P2P systems must model participants’ rationality, motivation, environmental friendliness, and conflicting interests while pursuing cost reduction, revenue maximization, and renewable-energy use.The paper identifies efficient and robust operation of heterogeneous, large-scale networks as a design requirement.
  • I. BACKGROUND AND MOTIVATION: Game theory is presented as an effective tool for modeling interactive and conflicting decision-making in P2P energy networks.The paper argues that P2P trading mechanisms require novel game-theoretic applications because central price signals may have limited influence.
  • I. BACKGROUND AND MOTIVATION: The paper systematically reviews game and auction approaches, introduces P2P features and an existing testbed, and examines selected trading models and their findings.It aims to help readers understand how game theory can be used in P2P energy trading and its potential benefits.

II. GAME THEORY FOR SMART ENERGY MANAGEMENT

Game theory provides a framework for analyzing strategic energy-management decisions whose outcomes depend on other participants’ actions. The paper organizes applications around non-cooperative and cooperative game concepts used across smart-energy domains.

  • II. GAME THEORY FOR SMART ENERGY MANAGEMENT: Game theory analyzes competitive situations in which each participant’s outcome depends on the actions chosen by other participants.It is divided into non-cooperative and cooperative branches.
  • II. GAME THEORY FOR SMART ENERGY MANAGEMENT: Non-cooperative games model strategic decisions among players with potentially conflicting interests, without requiring coordination or communication of strategic choices.The term describes how cooperation is enforced, not necessarily that players never cooperate.
  • II. GAME THEORY FOR SMART ENERGY MANAGEMENT: A non-cooperative game represents each player’s strategy and utility, with a player’s utility affected by the other players’ actions.The standard solution concept discussed is Nash equilibrium.
  • II. GAME THEORY FOR SMART ENERGY MANAGEMENT: Static games involve one action choice, whereas dynamic games involve repeated actions, time, and information about other players’ choices.Both types may use deterministic pure strategies or probabilistic mixed strategies.
  • II. GAME THEORY FOR SMART ENERGY MANAGEMENT: Nash equilibrium is a stable action profile where no player can improve utility by unilaterally changing its action, although pure-strategy existence is not guaranteed.Games may also contain multiple equilibria requiring selection of an efficient and desirable solution.

2) Cooperative game:

Cooperative games study incentives for independent decision makers to act together through coalitions, bargaining, or communication structures. The reviewed EV studies apply these and auction-based models to scheduling, pricing, storage sharing, and P2P trading.

  • 2) Cooperative game:: Cooperative games provide incentives for independent decision makers to act together and analyze how coalitions are structured.Coalitional games represent players and a value function quantifying each coalition’s worth.
  • 2) Cooperative game:: Canonical coalitional games study grand-coalition stability, coalition gains, and fair revenue allocation, with the core linked to coalition stability.Superadditivity means forming the grand coalition is not detrimental to players.
  • 2) Cooperative game:: Coalition-formation games account for cooperation costs and network structure, while dynamic versions study adaptation to changing players and topology.Because coalition gains are cost-limited, grand-coalition formation is described as rare.
  • 2) Cooperative game:: Coalitional graph games incorporate communication connectivity and seek distributed algorithms for building networks with stable and efficient structures.Communication structure can affect coalition utility and other game characteristics.
  • B. Energy Management in EV Domain: EV studies use Nash games and auctions for charging schedules, price competition, storage trading, and P2P electricity exchange.Reported outcomes include unique Nash equilibria, incentive compatibility, and at least one Nash equilibrium in auction mechanisms.

C. Energy Management in DER and Storage Domain

Game- and auction-theoretic approaches are reviewed for energy management with distributed energy resources and storage. The cited studies address consumer costs, storage sharing, direct trading, microgrid coalitions, and utility–user interactions.

  • C. Energy Management in DER and Storage Domain: Widespread DER deployment can support a clean and reliable energy system, but intermittent production creates integration challenges for the power system.Energy storage and management techniques are identified as ways to address intermittency and increase DER benefits.
  • C. Energy Management in DER and Storage Domain: Game-theoretic billing and optimization schemes schedule consumption, shift peak-time demand, reduce trading costs, and preserve user privacy with limited central communication.One distributed algorithm provides optimal production and storage strategies on users’ smart meters.
  • C. Energy Management in DER and Storage Domain: Auction mechanisms are used for renewable-energy and storage-capacity trading, including a realistically implemented reverse-auction model for microgrid operations.The implementation was reported as suitable for an existing electric utility grid.
  • C. Energy Management in DER and Storage Domain: An auction game for shared storage ownership has incentive compatibility and individual rationality, allowing residential units to choose the storage capacity fraction they share.The mechanism supports storage assistance for shared-facility controllers.
  • C. Energy Management in DER and Storage Domain: Coalitional and hierarchical games support direct DER-user trading, microgrid coalition formation, and utility–user benefit analysis.Reported results include a Shapley value in the core, a unique Stackelberg equilibrium, and reduced total cost for a shared facility controller.

D. Energy Management in Service Domain

Game-theoretic approaches have been applied extensively to energy-management services, including demand response, frequency regulation, storage sharing, and user incentives. The paper surveys these applications and notes that P2P-specific discussion remains limited.

  • Nash games: Nash games support demand-response services by scheduling users’ energy-related activities and can also form part of hierarchical games.The surveyed examples use Nash games alone or as follower-level interactions within Stackelberg frameworks.
  • Auction games: Auction games are used for storage sharing, demand response, and frequency regulation, including negotiations between aggregators and networked microgrid agents.One reverse-auction architecture models microgrid bidding behavior and short-term policy decisions to maximize individual profit.
  • Coalition games: Coalition games address demand-response regulation, direct community energy trading, and EV charging and discharging while preserving grid stability.The examples include coalition formation for EV scheduling and incentives for rooftop-solar users to trade within a community.
  • Hierarchical games: Hierarchical and Stackelberg games cover voltage and frequency regulation, demand response, storage sharing, and V2G coordination.A hierarchical Markov game optimizes an aggregator’s regulation capacity and strengthens its ability to bid a favorable frequency-regulation price.
  • Scope of the literature: Game theory is extensively used across smart-energy management, whereas its application specifically to P2P energy trading remains limited.The paper attributes this limited P2P discussion to the recent emergence and exploration of P2P trading.

III. GAME THEORY FOR ENERGY MANAGEMENT IN P2P NETWORK

A P2P energy network combines peer-controlled resource sharing with virtual market infrastructure and a physical distribution network. Its design requires coordinated grid connection, information access, market operation, pricing, participation, and automated energy management.

  • P2P energy network: P2P networks let members share resources, distribute energy among peers or storage, and control community production and consumption without central authority.The network includes consumers and prosumers with small-scale distributed energy resources such as rooftop solar and small wind turbines.
  • Network layers: The virtual layer securely transfers peer data, creates buy and sell orders, matches trades, processes payment, and coordinates energy exchange over the physical layer.The physical layer is the distribution grid, while the virtual platform provides the local electricity-market infrastructure.
  • Key features: A P2P network should define grid connection points, provide equal information access, and support market allocation, payment rules, bidding, and near-real-time order matching.Market operation may span day-ahead and intraday horizons and must respect generation-based allocation limits.
  • Grid connection: Island-mode operation requires a physical microgrid and sufficient participant generation capacity; trading solely over the traditional distribution network cannot provide it.The physical microgrid can decouple from the main grid during emergencies, whereas the traditional network alone cannot support island operation.
  • Pricing mechanism: Pricing mechanisms balance supply and demand through mechanisms such as individual- or uniform-clearing-price auctions, with surplus expected to lower trading prices.P2P pricing differs from traditional energy markets because renewable energy has zero marginal cost and lacks comparable taxes and surcharges.
  • Automatic energy management: Automatic energy-management systems use real-time demand and supply information to forecast profiles and develop bidding strategies under participant preferences and price limits.A rational user buys from the microgrid market when the price falls below its maximum price limit.

2) Brooklyn TransActive P2P project:

The Brooklyn TransActive P2P project combines a traditional-grid physical layer with a blockchain-based virtual market. Automated trading uses participant preferences, double auctions, predefined payments, and subsequent physical energy delivery.

  • Project architecture: The Brooklyn Microgrid project operates across three distribution grids in Brooklyn, New York, and includes a physical microgrid that can decouple during emergencies.The project is run by LO3 Energy and includes Borough Hall, Park Slope, and Bay Ridge participants.
  • Project architecture: Its virtual layer is separate from the physical layer and uses the TransActive blockchain architecture, with each peer requiring a blockchain account and meter.TransActive meters transfer household generation and demand data to the associated blockchain account.
  • Automated trading: An automatic energy-management system conducts most trading after participants specify energy-source preferences and price limits through the BMG App.Participants can change preferences at any time or retain one preference without further application interaction.
  • Market mechanism: The market uses a closed-order-book, time-discrete double auction with 15-minute time slots and merit-order allocation.Consumers bid maximum prices, prosumers bid minimum selling prices, and the last allocated bid sets the clearing price.
  • Settlement and delivery: After matching, financial transactions follow predefined payment rules in the virtual layer, while prosumers feed renewable generation into the distribution grid for consumers.The paper separates virtual-layer payment and trading from physical-layer electricity delivery.
  • Game-theoretic applications: The paper next examines auction, coalition, and hybrid auction–Stackelberg games across EV, DER and storage, and service domains.These examples are used to show how game-theoretic approaches can support different energy-management objectives.

B. P2P energy trading in EV domain

In EV P2P trading, local aggregators coordinate charging and discharging EVs through auction mechanisms designed around utility, cost, and social-welfare objectives. The reviewed auction approach seeks efficient allocation without requiring complete private information.

  • EV market structure: Charging, discharging, and idle EVs choose roles according to their current energy state, while local aggregators broker electricity and communication services.Aggregators collect local demand, receive selling prices from surplus EVs, and match trading pairs through iterative double auctions.
  • Utility and cost modeling: The EV model defines charging satisfaction and discharging cost functions that vary with energy price, charging willingness, and cost factors.The functions represent the real-valued benefit and cost obtained by participating EVs.
  • Social-welfare objective: A local aggregator selects charging and discharging energy vectors to maximize social welfare and allocation efficiency.The social-welfare objective is formed from the difference between participating EVs’ satisfaction and cost functions.
  • Information challenge: Social-welfare maximization requires complete information about EV utility and cost functions, including private states such as battery capacity.Because EVs may not share this information, the mechanism must extract hidden information from them.
  • Auction process: The auction mechanism iteratively collects bids, computes allocation vectors, updates optimal bids, and repeats until predefined optimality criteria are met.The auctioneer broadcasts allocations, benchmarks submitted optimal bids, and restarts the iteration when the criteria are not satisfied.
  • Reported properties: The auction-based approach obtains efficient energy allocation and optimal social welfare without requiring participants to share complete private satisfaction information.The double auction is described as individually rational and weakly budget balanced, supporting truthful bidding and preventing broker losses.

C. P2P Energy Trading in DER and Storage Domain

The paper models DER-and-storage P2P trading as a coalition game in which suppliers and end-users share benefits through a strategically determined trading price and revenue allocation. The scheme uses an asymptotic Shapley value to preserve fairness and stability as the customer population grows.

  • Coalition-game design: P2P trading sets pp2p between wholesale and retail prices, benefiting both small-scale sellers and end-users.The pricing condition is pwp ≤ pp2p ≤ prp.
  • Coalition-game design: The coalition game values participants’ net surplus and deficiency after internal P2P trades and residual market transactions.The value function accounts for surplus sold at pwp and deficiency purchased at prp.
  • Coalition properties: A non-empty core ensures that no subgroup can gain more by deviating, making the coalition stable when prp > pwp.Stability means customers continue participating in P2P trading.
  • Revenue allocation: Revenue is allocated using Shapley values so each customer’s payment corresponds to its contribution to P2P trading.The paper notes that exact Shapley-value computation becomes prohibitively expensive for very large coalitions.
  • Scalability: An asymptotic Shapley value lies within the coalition core, supporting stability and suitability for networks with very large numbers of customers.The paper presents this approximation as a response to the computational burden of exact Shapley values.

D. P2P Energy Trading in Service Domain

In the service domain, P2P trading shares residential storage with shared facility controllers through a modified auction. A Stackelberg-based payment rule determines a unique agreed price, while allocation rules distribute storage and oversupply burdens.

  • System model: The service-domain system connects residential units and shared facility controllers, each with generation and storage, to share storage capacity.Shared facilities include lifts, corridor lights, water pumps, and heat pumps.
  • Auction design: The modified auction uses determination, payment, and allocation rules to select participants, set payment, and assign shared storage.Participation depends on bids, offered and leased storage, and the Vickrey price.
  • Auction design: A Stackelberg game sets the auction price by balancing SFC cost savings and storage needs with residential units’ storage-sharing benefits.The auctioneer chooses price, while residential units choose storage capacity to offer.
  • Equilibrium: The Stackelberg game always has a unique solution, producing an equilibrium price that all residential units and SFCs accept without deviation.The price governs energy-storage sharing between the two participant groups.
  • Allocation: When storage supply exceeds SFC demand, oversupply burden is distributed through either proportional or equal allocation.Proportional allocation uses reservation prices; equal allocation distributes burden equally among participating residential units.
  • Incentives and scope: The payment and proportional-allocation rules are individually rational, and these properties persist in the time-varying extension.The resulting auction is incentive compatible, so participants have no incentive to cheat.

IV. OUTCOMES OF GAME THEORETIC APPLICATIONS IN P2P ENERGY NETWORK

The reviewed studies show that game-theoretic P2P designs can improve cost reduction, utility, and energy utilization across EV, DER, storage, and service applications. Outcomes depend on pricing, transmission losses, market size, and generation mix.

  • Cross-domain findings: The review covers game-theoretic P2P applications in EV, DER-and-storage, and service domains, comparing their outcomes with existing schemes.It frames the results as evidence of P2P and game theory’s effectiveness for network benefits.
  • EV domain: When the grid’s sell-out price is below local EV sellers’ price, hybrid trading benefits buyers through a lower average buying price.Higher transmission losses still require more electricity to meet the same demand.
  • EV domain: The proposed P2P EV model has lower energy loss and higher electricity-utilization efficiency than the hybrid model.The paper attributes the efficiency difference to lower transmission loss.
  • EV domain: EV participation depends on pricing: the average P2P buying price is 0.17 dollar/kWh, and the scheme attracts EVs during peak-price periods.The paper states that the buying price should be lowered to compete with the hybrid market.
  • EV domain: P2P selling prices also require careful selection because rising hybrid-market prices can shift EV sellers toward the smart grid.Mid-rate pricing is offered as one suitable example.

B. DER and Storage Domain

The DER-and-storage results indicate that P2P trading can generate revenue for suppliers and savings for end-users, while optimal generation mixes vary with market size. Supplier revenue falls as supply expands, whereas an appropriate solar–wind mix can maximize total revenue.

  • Revenue outcomes: P2P participants can reach monthly revenues of 80 dollars for suppliers and 62 dollars for end-users.The comparison reports a 110-dollar monthly bill for a household without P2P and savings of up to 60%.
  • Market-size effects: Supplier revenue decreases as the number of suppliers increases because greater available supply lowers the trading price.The paper links this market effect to increased energy available for sale.
  • Generation mix: When fewer than 24 suppliers participate, buyers and sellers prefer solar; above 72, both prefer wind.Between 24 and 72 suppliers, end-users prefer solar while suppliers prefer wind, making mixed generation advisable.
  • Generation mix: For 30 suppliers, total revenue is maximized when 60% of suppliers use solar generators.The optimal solar percentage approaches zero as the number of suppliers increases.
  • Implications: The paper concludes that participant cooperation can benefit both sides by enabling internal trading without relying on less-attractive main-grid pricing.It also notes that the benefit depends on how the pricing scheme is designed.

C. Service Domain

The service-domain analysis examines P2P sharing of residential storage for SFCs, showing how storage demand and reluctance shape residential utility and comparative performance. Utility rises with SFC storage requirements until shared capacity saturates, while the auction scheme outperforms ED and FiT benchmarks.

  • Service-domain setup: Each residential unit is modeled as a group of 5–25 households equipped with a 25 kWh storage device, while each SFC requires 100–500 kWh.The required storage may differ when users’ usage patterns change.
  • Service-domain setup: The study models joint energy-storage sharing among residential units and SFCs through a P2P trading and auction scheme.Residential units place available storage into the market, where it is allocated according to the scheme’s allocation rule.
  • Demand and utility: Residential utility initially increases with SFC storage demand, then stabilizes once each unit reaches its maximum shareable storage capacity.Higher demand allows more reserved storage to be shared, but additional demand cannot increase utility after shared capacity saturates.
  • Reluctance effects: Lower reluctance corresponds to greater willingness to share, favoring lower-reluctance units when SFC demand is high and higher-reluctance units when demand is low.Lower-reluctance units share more storage and gain more at higher demand, whereas higher-reluctance units bear less oversupply burden at lower demand.
  • Comparative performance: Performance improvement is more pronounced as storage demand rises from 200 to 350, but becomes less noticeable when demand increases from 400 to 450.Once residential storage spaces reach saturation, higher SFC demand does not change the amount shared.
  • Key insights: The summarized findings indicate that low-demand SFCs benefit from higher-reluctance residential units, whereas high-demand SFCs benefit from lower-reluctance units.The P2P scheme is more beneficial for residential units than ED and FiT sharing in the considered cases.

V. CONCLUSION

The paper surveys game-theoretic approaches for P2P energy management, organizing prior work across energy domains and examining selected models in detail. It identifies consumer participation, grid benefits, security, incomplete information, and physical-network constraints as important directions for future development.

  • The paper provides an overview of game-theoretic approaches for energy management and divides its discussion into EV, DER and storage, and service domains.
  • It gives detailed discussions of specific game-theoretic approaches in each P2P network domain and summarizes the interpretation of their outcomes.
  • P2P energy-management research remains relatively new, with platforms still in the pilot phase before integration into current energy systems.
  • Consumer-centric P2P schemes should provide participant benefits because some recent energy-trading models and pilots were discontinued after consumers rejected them.
  • P2P trading should demonstrate benefits to the grid, which may participate as a leader or follower in Stackelberg-game models.
  • Blockchain can address security and privacy concerns, but integrating it with game theory must account for its potentially high computational power requirements.
  • Future P2P energy-trading models should address incomplete information and incorporate Kirchhoff-law coupling constraints, which complicate market analysis and design.
Loading 1804.00962v1…