Comments on: Computation of Correlated Equilibira (cont.)
https://agtb.wordpress.com/2010/11/02/computation-of-correlated-equilibira-cont/
Computation, Economics, and Game TheorySat, 01 Jan 2011 06:49:12 +0000
hourly
1 http://wordpress.com/
By: AGT/E 2010 year in review « Algorithmic Game-Theory/Economics
https://agtb.wordpress.com/2010/11/02/computation-of-correlated-equilibira-cont/#comment-4208
Sat, 01 Jan 2011 06:49:12 +0000http://agtb.wordpress.com/?p=1060#comment-4208[…] On the core algorithmic/complexity questions of the computational difficulty of finding equilibria in games and markets we have seen only little progress, mostly for some market variants. There was some commotion regrading the complexity of correlated equilibria in concise games, but that was soon settled. […]
]]>
By: Albert Jiang
https://agtb.wordpress.com/2010/11/02/computation-of-correlated-equilibira-cont/#comment-2664
Fri, 05 Nov 2010 02:40:57 +0000http://agtb.wordpress.com/?p=1060#comment-2664Thanks for the very nice summary of our result! In reply to the comment by Jack, we believe that our approach has some independently interesting properties, regardless of the severity of the issues in the Papadimitriou&Roughgarden paper. For example, the resulting correlated equilibrium has polynomial-sized support, which is simpler to represent and to sample from than the mixture of product distributions generated by the previous algorithms. Also, while our approach for computing a small-support correlated equilibrium requires an efficient subroutine for computing expected utility under product distributions, *verifying* such a small-support correlated equilibrium does not require such a subroutine.
]]>
By: Jack
https://agtb.wordpress.com/2010/11/02/computation-of-correlated-equilibira-cont/#comment-2661
Tue, 02 Nov 2010 14:01:15 +0000http://agtb.wordpress.com/?p=1060#comment-2661I’m confused — based on the comments in your last post, it seemed like there were not actually serious problems with the Papadimitriou/Roughgarden paper, and that the MIT authors withdrew their paper. Is the original paper flawed, or not?
]]>