1: \begin{abstract}

2: We consider the problem of distributed channel allocation in large

3: networks under the frequency-selective interference channel. Performance

4: is measured by the weighted sum of achievable rates. Our proposed

5: algorithm is a modified Fictitious Play algorithm that can be implemented

6: distributedly and its stable points are the pure Nash equilibria of

7: a given game. Our goal is to design a utility function for a non-cooperative

8: game such that all of its pure Nash equilibria have close to optimal

9: global performance. This will make the algorithm close to optimal

10: while requiring no communication between users. We propose a novel

11: technique to analyze the Nash equilibria of a random interference

12: game, determined by the random channel gains. Our analysis is asymptotic

13: in the number of users. First we present a natural non-cooperative

14: game where the utility of each user is his achievable rate. It is

15: shown that, asymptotically in the number of users and for strong enough

16: interference, this game exhibits many bad equilibria. Then we propose

17: a novel non-cooperative M Frequency-Selective Interference Channel

18: Game (M-FSIG), as a slight modification of the former, where the utility

19: of each user is artificially limited. We prove that even its worst

20: equilibrium has asymptotically optimal weighted sum-rate for any interference

21: regime and even for correlated channels. This is based on an order

22: statistics analysis of the fading channels that is valid for a broad

23: class of fading distributions (including Rayleigh, Rician, m-Nakagami

24: and more). We carry out simulations that show fast convergence of

25: our algorithm to the proven asymptotically optimal pure Nash equilibria.

26: \end{abstract}