1: \begin{abstract}

2: A major goal in Algorithmic Game Theory is to justify equilibrium concepts from an algorithmic and complexity perspective.

3: One appealing approach is to identify natural distributed algorithms that converge quickly to an equilibrium.

4: This paper established new convergence results for two generalizations of Proportional Response in Fisher markets with buyers having CES utility functions.

5: The starting points are respectively a new convex and a new convex-concave formulation of such markets.

6: The two generalizations correspond to suitable mirror descent algorithms applied to these formulations.

7: Several of our new results are a consequence of new notions of strong Bregman convexity

8: and of strong Bregman convex-concave functions, and associated linear rates of convergence, which may be of independent interest.

9: 

10: Among other results, we analyze a damped generalized  Proportional Response and show a linear rate of convergence in

11: a Fisher market with buyers whose utility functions cover the full spectrum of CES utilities aside the extremes of linear and Leontief utilities;

12: when these utilities are included, we obtain an empirical $O(1/T)$ rate of convergence.

13: 

14: \smallskip

15: 

16: \noindent\bold{Keywords.} Proportional Response, mirror descent, Bregman divergence, Fisher market

17: \end{abstract}

18: