Communication complexity of Nash equilibrium in potential games

Yakov Babichenko,Aviad Rubinstein

DOI: https://doi.org/10.48550/arXiv.2011.06660

2020-11-13

Abstract:We prove communication complexity lower bounds for (possibly mixed) Nash equilibrium in potential games. In particular, we show that finding a Nash equilibrium requires $poly(N)$ communication in two-player $N \times N$ potential games, and $2^{poly(n)}$ communication in $n$-player two-action games. To the best of our knowledge, these are the first results to demonstrate hardness in any model of (possibly mixed) Nash equilibrium in potential games.

Computer Science and Game Theory,Computational Complexity

What problem does this paper attempt to address?

This paper attempts to address the communication complexity problem of finding Nash equilibria (including mixed strategy Nash equilibria) in potential games. Specifically, the authors prove that in a two-player \(N \times N\) potential game, the communication complexity required to find a Nash equilibrium is at least \(\Omega(N^c)\), where \(c > 0\) is a constant; while in a potential game with \(n\) players and two actions, the communication complexity is at least \(2^{\Omega(\sqrt{n})}\).

These results are the first to demonstrate the difficulty of finding Nash equilibria in potential games in any model. Prior to this, although there were some results on the difficulty of pure strategy Nash equilibria, these results did not apply to mixed strategy Nash equilibria. Therefore, this paper fills this gap by proving that even mixed strategy Nash equilibria are difficult to compute in potential games.

### Main Contributions of the Paper:
1. **Lower Bound on Communication Complexity**: The authors prove that in a two-player \(N \times N\) potential game, the communication complexity required to find a Nash equilibrium is at least \(\Omega(N^c)\).
2. **Lower Bound on Communication Complexity in Multi-Player Games**: In a potential game with \(n\) players and two actions, the communication complexity is at least \(2^{\Omega(\sqrt{n})}\).
3. **Extension to Congestion Games**: These results not only apply to potential games but also extend to congestion games, further proving the lower bound on communication complexity for finding Nash equilibria in congestion games.

### Technical Methods:
1. **Simulation Theorem**: The authors use the "simulation theorem" to elevate the results from query complexity to communication complexity.
2. **Potential Function Construction**: The authors construct a complex potential function that embeds the discrete "End-of-Line" problem into a continuous potential function, such that the local maximum corresponds to the end of the line.
3. **Mixed Strategy Analysis**: By introducing additional incentive mechanisms, the authors ensure that the information reported by players is truthful and that a Nash equilibrium can be found under mixed strategies.

### Open Problems:
1. **Fine-Grained Communication Complexity**: Closing the gap between \(\Omega(N^c)\) and \(\tilde{O}(N)\) in two-player games, and between \(2^{\Omega(\sqrt{n})}\) and \(2^{O(n)}\) in multi-player two-action games.
2. **Computational Complexity**: Further investigating the computational complexity of related problems in PPAD∩PLS.

### Summary:
Through rigorous mathematical proofs, this paper demonstrates the lower bounds on communication complexity for finding Nash equilibria in potential games, filling a gap in this field. These results are not only significant for theoretical computer science but also provide important references for practical applications in game theory and economics.