⚠️ Warning: This is a draft ⚠️

This means it might contain formatting issues, incorrect code, conceptual problems, or other severe issues.

If you want to help to improve and eventually enable this page, please fork RosettaGit's repository and open a merge request on GitHub.

==Suggestions for an improved task description== I suggest to add a second usage case, to be more sure the code entries are correct. -[[User:Bearophile|bearophile]] ([[User talk:Bearophile|talk]])

An external reference to the algorithm might be good too; as well as an explanation of what VAM stands for.

It might be best to change the name of the task to: "Vogel's approximation method (VAM)". --[[User:Paddy3118|Paddy3118]] ([[User talk:Paddy3118|talk]]) 03:43, 4 September 2013 (UTC)

:I have added a reference. I have also changed the task with a rule as to which row or column to choose when tied. This is probably better for an RC task as all tasks will get the same answer. At the moment it seems to depend on how max works in the particular language. It will also be much more fun to program! The change is a simplification of one of the improvements to VAM discussed in http://www.mcajournal.org/articleinpress/articleinpress_955.pdf. This paper is for further study (i.e. I've only read it quickly). It includes another example we could use (Table 2). Section 4.2 Experimental design discusses a testing method we could use, generating a number of random tables and solving them using VAM and GLPK and comparing the two. Perhaps we don't need 12000.--[[User:Nigel Galloway|Nigel Galloway]] ([[User talk:Nigel Galloway|talk]]) 11:09, 4 September 2013 (UTC)