Consumer Flexibility Aggregation Using Partition Function Games With Non-Transferable Utility

Tiago Pinto, Michael J. Wooldridge, Zita A. ValeIEEE Access 2021 · OpenReview

Abstract

This paper explores the aggregation of electricity consumers' flexibility through a novel coalitional game theory model for partition function games with non-transferable utility. This model formalizes a game where electricity consumers form coalitions to trade their consumption flexibility in the electricity market. Utility functions are defined to measure player preferences, balancing income, internal competition, and market externalities. The model was evaluated using a simple illustrative case and a large-scale scenario with over 20,000 consumers based on real-world data. Results demonstrate that the proposed model achieves solutions more suitable for consumers than traditional aggregation techniques, such as clustering. By considering the perspectives of all involved players, the model reflects rational behavior rather than imposing solutions beneficial only to the aggregator.

Introduction

As power grids increasingly integrate variable renewable energy sources, "demand-side flexibility"—the ability of consumers to shift or reduce electricity usage—has become a critical tool for grid balancing. However, individual residential and small commercial consumers lack the necessary volume to participate directly in wholesale energy markets. Consequently, they must be grouped by "aggregators."

Historically, researchers have relied on top-down clustering algorithms like K-Means to group similar consumers. While efficient for the aggregator, these methods often ignore the rational, self-interested nature of consumers. Grouping similar consumers can lead to high internal competition when flexibility is called upon. Furthermore, traditional models often overlook "externalities," such as how the total number of market participants affects final prices. This research proposes a bottom-up aggregation model designed to align with individual consumer preferences and market realities.

Methodology

The authors frame the aggregation problem as a Coalitional Game, specifically utilizing a Partition Function Game with Non-Transferable Utility (NTU).

  • Partition Function: Unlike standard games, this accounts for externalities, recognizing that a coalition's payoff depends on how the rest of the market is structured (e.g., more competing coalitions can drive down prices).
  • Non-Transferable Utility (NTU): Since players have vastly different baseline flexibilities (e.g., an industrial plant versus a household), utility cannot be easily redistributed as a single pool of money. NTU respects these distinct scales and preferences.
Proposed game-theoretic framework for coalition formation and social welfare maximization.

The model employs a custom utility function for each player that balances four primary objectives: maximizing expected income, maximizing their relative volume of flexibility within their coalition, minimizing the total number of competing coalitions in the market, and maximizing the diversity of players within their own coalition to reduce internal competition. To navigate the exponentially large search space of possible coalition structures, the authors use a Genetic Algorithm (GA) to find near-optimal groupings that maximize social welfare.

Evaluation and Key Insights

The model was tested against two scenarios: a 3-player illustrative case and a large-scale real-world scenario involving 20,310 consumers in a Portuguese distribution network. The real-world data included five consumer categories (domestic, small/medium/large commerce, and industrial) with actual flexibility volumes and pricing.

Key Findings

  • The Flaw of Clustering: Traditional clustering groups similar players, which maximizes internal competition from a consumer's perspective. The proposed model shows that consumers actually prefer diverse coalitions where their specific flexibility profile is unique.
  • Superior Social Welfare: In the large-scale scenario, the game-theoretic model achieved Utilitarian and Egalitarian Social Welfare scores more than twice as high as standard clustering algorithms (including K-Means and meta-heuristic approaches like ADE-LPC).
  • Incentive for Grand Coalitions: The model proves that consumers are incentivized to form large, diverse groups to minimize market competition, thereby protecting their profit margins.

Conclusion and Future Work

While the proposed model offers significantly better alignment with consumer interests and higher social welfare, it comes with a higher computational cost. The Genetic Algorithm required approximately two hours to process 20,000+ players, compared to minutes for simple clustering. However, for strategic market planning, this trade-off is considered acceptable.

Future research will explore distributed computing to reduce the computational burden and incorporate advanced fairness metrics, such as the Shapley value, to further refine the model's ability to predict and incentivize rational player behavior.

Summary updated

Abstract

This paper explores the aggregation of electricity consumers flexibility. A novel coalitional game theory model for partition function games with non-transferable utility is proposed. This model is used to formalize a game in which electricity consumers find coalitions among themselves in order to trade their consumption flexibility in the electricity market. Utility functions are defined to enable measuring the players preferences. Two case studies are presented, including a simple illustrative case, which assesses and explains the model in detail; and a large-scale scenario based on real data, comprising more than 20,000 consumers. Results show that the proposed model is able to reach solutions that are more suitable for the consumers when compared to the solutions achieved by traditional aggregation techniques in power and energy systems, such as clustering-based methodologies. The solutions found by the proposed model consider the perspectives from all players involved in the game and thus are able to reflect the rational behaviour of the involved players, rather than imposing an aggregation solution that is only beneficial from the perspective of the aggregator.
BibTeX
@inproceedings{pinto2021consumer,
  author = {Tiago Pinto and Michael J. Wooldridge and Zita A. Vale},
  title = {{Consumer Flexibility Aggregation Using Partition Function Games With Non-Transferable Utility}},
  booktitle = {IEEE Access},
  year = {2021},
  url = {https://openreview.net/forum?id=X75d6nEEZL},
}
Figure