Comments on: Information Geometry (Part 12)

By: Network Theory for Economists « Azimuth

Network Theory for Economists « Azimuth — Tue, 15 Jan 2013 23:28:47 +0000

In this talk I explain how reaction networks can be used to describe evolutionary games. I point out that in certain classes of evolutionary games, evolution tends to increase ‘fitness’ [...]

By: Information Geometry (Part 13) « Azimuth

Information Geometry (Part 13) « Azimuth — Tue, 26 Jun 2012 16:33:29 +0000

[...] Last time I gave a sketchy overview of evolutionary game theory. Now let’s get serious. [...]

By: John Baez

John Baez — Tue, 26 Jun 2012 07:19:57 +0000

People have studied a variety of different equations of motion in evolutionary game theory: • Samuel Alizon and Daniel Cownden, Replicator dynamics. Some of these describe agents that look at what other agents are doing and copy them if those other guys are doing better than they are.

By: Marc Harper

Marc Harper — Tue, 26 Jun 2012 00:28:52 +0000

Ah, my bad. I should have looked more closely (I assumed it was the original paper that was linked).

By: John Baez

John Baez — Mon, 25 Jun 2012 17:20:12 +0000

In economics people use evolutionary game theory to model a population of people who adapt their mixed strategies to do well, instead of a population of replicators each with a pure strategy, who evolve to do well. Under some assumptions the two math problems are isomorphic. So I think this work could still fall under the heading of evolutionary game theory, though I’m not sure.

By: Nathan Urban

Nathan Urban — Mon, 25 Jun 2012 16:06:09 +0000

The comment by Sigmund & Nowak at the end of the Edge article points out that the Press & Dyson paper isn’t really about evolutionary game theory. Evolutionary game theory studies games on evolving populations of players. P&D only study two-player (iterated) games. Their definition of “evolutionary” really just means “adaptive”: “We can say, loosely, that is an evolutionary player if he adjusts his strategy according to some optimization scheme designed to maximize his score , but does not otherwise explicitly consider X’s score or her own strategy.”

By: John Baez

John Baez — Mon, 25 Jun 2012 09:36:42 +0000

I linked to that in my blog post. Yes, everyone should check it out! The commentary starts out being a bit vague and fuzzy, but gets more precise as it goes along.

By: davidtweed

davidtweed — Mon, 25 Jun 2012 08:59:38 +0000

Also, although it's not obvious, if you click through that commentary and then go to the "Full text" sidebar, the original paper is actually openly downloadable.

By: Marc Harper

Marc Harper — Mon, 25 Jun 2012 05:20:22 +0000

Edge.org has an interesting discussion of the result by Press and Dyson here: http://edge.org/conversation/on-iterated-prisoner-dilemma