Open 8ac3f0f3-bd13-47e7-baee-523ad1646342 opened 9 years ago
Commit: 1910055
Branch: u/ptigwe/dep_maximization
Dependencies: #18536
Just wondered why this ticket depends on 18536? Otherwise I'm happy to start reviewing it.
I had two main options: branch off from #18536 and deprecate it in all functions including the LP solvers; and the second was to deprecate both within this ticket and #18536. I chose the first as it places the deprecation in one ticket.
Actually, come to think of it, I could have just deprecated within this ticket and removed maximization from the LP solvers completely.
You could probably still do that... You could comment on #18536 to check if Karl has started reviewing and if not set it back to 'needs work'...
You could probably still do that... You could comment on #18536 to check if Karl has started reviewing and if not set it back to 'needs work'...
I won't be looking at any of this today, anyway.
I won't be looking at any of this today, anyway.
Perfect :)
Branch pushed to git repo; I updated commit sha1. Last 10 new commits:
60efdc7 | Fixed '_as_gambit_game' to support 'maximization' parameter |
0e300ae | Fixed indentation and removed incorrect error |
313f42c | Updated tests for cbc and PPL |
a24c7dd | Included tests for constant-sum non-zero sum game and included maximization in the LP solver |
2f44485 | Tests for single / multiple Nash equilibria |
fb4461c | Fixed minor error with LP solver |
e4107dc | Updated tests for normal form games |
c225b92 | Remove maximization from LP functions as it is going to be deprecated |
93229b7 | Revert "Remove maximization from LP functions as it is going to be deprecated" |
73b5780 | Deprecate maximization from Normal Form Games |
There are some changes which were made in the #18536 ticket which would have caused a few merge conflicts. So in order to avoid that, I think it's best to stick with the current setting.
Which order should these tickets be reviewed in, if any?
Branch pushed to git repo; I updated commit sha1. New commits:
6e2aae5 | Merge branch 'develop' into gt_extension |
92345cc | Modified the '_gambit_' function to support n-player games |
2c6aee7 | Update doctests |
06d6b4c | Updated doctests of `catalog` to use `enumeration` |
122f8e1 | Merge branch 'gt_extension' into dep_maximization |
The branch no longer merges with current sage releases.
This ticket would be for the deprecation of the
maximization
parameter within theNormalFormGame
class which is used be various functions when computing the Nash equilibrium. Some of the reasons include:maximization=False
is equivalent to solving an instance with negative payoffs. So in the two player instance, one might as well create an instance of (-A, -B) that way, creating the same instances over and over when you want to compute equilibria using different algorithms.Depends on #18536
CC: @drvinceknight @dimpase @nathanncohen @kcrisman
Component: game theory
Author: Tobenna P. Igwe
Branch/Commit: u/ptigwe/dep_maximization @
122f8e1
Issue created by migration from https://trac.sagemath.org/ticket/18679