Distributed Nash Equilibrium Seeking in Coalition Games for Uncertain Euler-Lagrange Systems With Application to USV Swarm Confrontation

Cheng Yuwen, Guanghui Wen, Jialing Zhou, Meng Luan, Tingwen Huang

TL;DR

This paper introduces a novel distributed algorithm for multiple uncertain Euler-Lagrange (EL) systems to find a Nash Equilibrium (NE) in coalition games. The approach handles unknown disturbances and both local and coupling constraints, a significant advancement for real-world multi-agent systems. It's demonstrated through a realistic USV swarm confrontation scenario, showcasing its practical applicability.

What Problem Does This Paper Address?

The paper tackles the complex challenge of achieving stable coordination (Nash Equilibrium) in multi-agent systems that operate under uncertain conditions and have inherent physical dynamics. Specifically, it focuses on "coalition games," where groups of agents (coalitions) cooperate internally to minimize a shared cost while competing against other coalitions. Existing methods often assume perfect knowledge of system dynamics, known disturbance bounds, or simplified agent models (like first-order systems). This paper addresses a more realistic scenario:

  • Uncertainty: Agents have Euler-Lagrange (EL) dynamics, which are common for mechanical systems, but these dynamics are uncertain.
  • Unknown Disturbances: External disturbances affect the agents, and their bounds are unknown.
  • Complex Constraints: Agents face both individual (local) constraints and constraints that link them to other agents within or across coalitions (coupling constraints).
  • Limited Information: Each agent only knows its own state, cost function, and constraint parameters, not the full state of all other agents.

This scenario is crucial for applications like robotic swarms, where agents must coordinate effectively despite imperfect information and unpredictable environments.

Methods

The authors propose a distributed NE seeking strategy that integrates several key techniques:

  • Adaptive Control for EL Systems: To handle uncertain EL dynamics and unknown disturbances, adaptive control methods are combined with sign functions. This allows the system to compensate for model uncertainties and suppress disturbances without needing to know their exact bounds.
  • Distributed Primal-Dual Dynamics: To manage both local and coupling constraints, primal-dual techniques are employed. Adaptive methods update Lagrange multipliers for local constraints, while consensus protocols help resolve multipliers for coupling constraints.
  • Gradient and Action Estimation:
    • Dynamic Average Consensus: This is used to estimate the gradient of each coalition's cost function.
  • Leader-Following Protocol: This is employed to estimate the actions of other agents, overcoming the limitation of not having global state information.
  • Lyapunov Stability Analysis: The theoretical framework relies on Lyapunov stability theory to prove that the proposed distributed algorithm guarantees global convergence to the Nash Equilibrium under standard assumptions (convexity, graph connectivity).

Datasets

Not discussed in the paper. The paper focuses on theoretical development and simulation of a custom scenario.

Key Insights and Ideas

  • Bridging Theory and Reality: The paper significantly advances the state-of-the-art by incorporating realistic complexities like uncertain EL dynamics, unknown disturbances, and both local and coupling constraints into NE seeking for coalition games. This makes the theoretical results more applicable to real-world multi-agent systems.
  • Decentralized Information Flow: The proposed strategy achieves NE seeking with agents only having access to local information, which is a critical requirement for many distributed systems.
  • Unified Framework for Complex Tasks: The application to USV swarm confrontation demonstrates the versatility of the coalition game framework. It allows for modeling diverse and adversarial tasks (attack, defense, formation, interception) within a single, coherent structure.
  • Robustness to Unknowns: The adaptive and sign-function-based approach provides robustness against unknown system parameters and external disturbances, a key challenge in practical deployments.

Key Results

  • Guaranteed Convergence: The paper theoretically proves that the proposed distributed algorithm ensures global convergence of the closed-loop system to the Nash Equilibrium under standard assumptions.
  • USV Swarm Simulation: A comprehensive simulation of a USV swarm confrontation scenario was conducted. This simulation involved:
  • Modeling of red and blue USV swarms with intrusive and defensive roles.
  • Formulation of diverse tasks: attack, interception, formation, and geometric center alignment.
  • Inclusion of spatial boundaries and logistical supply line constraints.
  • Demonstration of the algorithm's effectiveness in achieving coordinated behavior and stability within the adversarial context.
  • Lagrange Multiplier Convergence: Simulations show that Lagrange multipliers for local constraints asymptotically approach zero, and those for coupling constraints also converge, indicating successful constraint satisfaction.

Limitations and Future Work

  • Computational Complexity: While distributed, the algorithm involves multiple estimation and update steps, which could lead to significant computational load on individual agents, especially in large swarms.
  • Parameter Tuning: The performance of the algorithm relies on selecting sufficiently large gains for the adaptive and consensus components (α, β, γ, κ). Tuning these parameters for optimal performance in complex scenarios might be challenging.
  • Future Work: The authors suggest developing distributed online NE seeking methods for time-varying coalition games. This would address scenarios where tasks, agent capabilities, or environmental conditions change dynamically over time.

Summary updated

Abstract

In this paper, a coalition game with local and coupling constraints is studied for uncertain Euler-Lagrange (EL) systems subject to disturbances with unknown bounds. In the coalition game, each agent collaborates with other agents within the same coalition to optimize its coalition's cost function while simultaneously competing against agents in other coalitions. Under a distributed framework where each agent has access only to its own action, cost function, and constraint parameters, a distributed strategy is proposed to seek the Nash equilibrium (NE). By combining adaptive methods and sign functions, model uncertainties and disturbances with unknown bounds in the EL system are compensated and suppressed, respectively. Furthermore, an integration of adaptive methods and consensus protocols is employed to update the Lagrange multipliers of both local and coupling constraints. A dynamic average consensus is employed to estimate the gradient of coalition function, while the leader-following protocol is adopted to estimate the actions of other agents. By leveraging Lyapunov theory, the NE is proven to be asymptotically stable. Moreover, an unmanned surface vehicle swarm confrontation is meticulously modeled and analyzed in the coalition game framework. A diverse array of tasks, including formation, encirclement, and interception, are systematically formulated. A numerical example demonstrates the effectiveness of the proposed algorithm.
BibTeX
@misc{yuwen2025distributed,
  author = {Cheng Yuwen and Guanghui Wen and Jialing Zhou and Meng Luan and Tingwen Huang},
  title = {{Distributed Nash Equilibrium Seeking in Coalition Games for Uncertain Euler-Lagrange Systems With Application to USV Swarm Confrontation}},
  year = {2025},
  eprint = {2504.04475},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2504.04475},
}
Figure