Talk:Price of anarchy

Not copyrighted

Note: Similarity to the GT CS course notes page is not a coincidence. The notes are not copyrighted, and were designed for the purpose of forming an introductory article to the price of anarchy for wikipedia, as well as a class resource.

It's too bad they didn't copy the figure as well. It is referenced but not shown. The figure is here --Michael Daly (talk) 05:17, 23 September 2008 (UTC)[reply]

I have tried to add some introductory points and move some math stuff under specific 'math section' headings. I feel this article is way too technical and needs still more editing to make it comprehensible to a lay person.Jasemurphy (talk) 23:40, 4 January 2010 (UTC)[reply]

I changed the text in the beginning and the mathematical section (also to discuss Price of Stability, games with cost function instead of welfare function and make the definition somewhat more general. Also added a simples Prisoner's Dilemma example. I think we still need to work on the work on the two CS-ish examples to make them more Wikipedia-like and less like lecture notes. Besides, does the intro looks comprehensible to a lay person? I think I improved it in that regard, but comments would be appreciated. Renatoppl (talk) 13:42, 15 September 2010 (UTC)[reply]

Added a link to https://wiki.cc.gatech.edu/theory/index.php/Price_of_anarchy which is plainly the source for this article (see the original CorenSearchBot warning & M. Daly's comment above). Athenray (talk) 10:19, 10 June 2011 (UTC)[reply]

It appears that no one pays attention to the talk in this section as problems were pointed out six years ago, and they remain. The section on Braess Paradox refers to a nonexistent figure. The set of linear functions L, in the same section, has each of the functions \ell_e indexed by the edges e\in E, but in fact these function do not depend on the edges. Each function only depends on the constants a and b. This is not completely wrong, mathematically speaking, but I would say that it is mathematically clutzy. (John B. October 2, 2014) — Preceding unsigned comment added by Johnbaillieul (talk • contribs) 23:26, 2 October 2014 (UTC)[reply]

Is it "selfish" behaviour or "uncoordinated" behaviour?

The opening sentence currently says "...measures how the efficiency of a system degrades due to selfish behavior of its agents". Is selfish really the right idea here? It seems from the formal definition that the concept is not about selfish-vs-selfless, but about coordinated-vs-uncoordinated. (It's possible to have a system in which selfish agents share information, or in which selfless agents cannot share information - so the distinction is important.) Can someone clarify which is the actual core of this concept? Then we can tighten up the writing in this article.--mcld (talk) 15:00, 5 March 2018 (UTC)[reply]