Talk:Transportation problem: Difference between revisions
→Task Description
mNo edit summary |
|||
| (6 intermediate revisions by 2 users not shown) | |||
Line 10:
* Maybe should add a more detailed description of the algorithm of solving this problem? [[User:Русский|Русский]] ([[User talk:Русский|talk]]) 09:45, 9 February 2014 (UTC)
: That would be nice - right now, the implementations here seem to be messy and overly complicated. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 00:54, 4 August 2015 (UTC)
:: Have you read the page at the first link I provided? It's not very long. You'll see it's not as simple as you think. After an initial estimate, you need to find the optimal solution, that's where the catch is. [[User:Fwend|Fwend]] ([[User talk:Fwend|talk]]) 19:12, 13 July 2016 (UTC)
::: Why is this even a question?
::: If you can't put the algorithm description on the task page, and if you can't include task examples which illustrate whatever issues concern you, then maybe the task doesn't belong here on this site. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 19:27, 13 July 2016 (UTC)
:::: Yeah, the task description is not well-phrased, I've been thinking about improving it (I didn't write it) but I looked into this task almost 2 years ago, so it's no longer clear in my mind. For the time being, the link I provided is your best bet. [[User:Fwend|Fwend]] ([[User talk:Fwend|talk]]) 19:33, 13 July 2016 (UTC)
::::: After re-reading http://orms.pef.czu.cz/text/transProblem.html which is what is currently linked in the task description. I am not convinced I need to make any changes to my stupid implementation. The northwest corner method is about coming up with an initial feasible solution, so that other methods may be used, and I'm using that. Presumably the "method of potentials (with redistributive cycle)" would involve additional work, but for the task example there is no additional work to be performed. Also, there's no description of this method at that link. (And, yeah, I sort of assumed I had already conveyed this idea, but since the task author(s) presumably sort of assumed that they had adequately described the task, I figured that if I can expect more of them I should also expect more of myself.)
::::: So, just to re-iterate: task needs a better example. Task needs a better description.
::::: And, for example, costs of 4 6 8,: 4 2 10 would probably be better than what's currently listed. But an example with more than two suppliers, more than two consumers would be a much better test of this kind of code. If possible, also, we should have a local optimum cost near the values supplied by the northwest method which is not the global optimum.
::::: Anyways... for now, the task itself is defective and I think we need at least one implementation which clearly illustrates this defect. And, since no one else has volunteered to be the village idiot, I am electing myself to that position. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 18:43, 15 July 2016 (UTC)
::::: That said, the [[Brace expansion]] task might be a good example of for how a decent task specification should look? --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 12:44, 20 July 2016 (UTC)
==Task Description==
Line 15 ⟶ 35:
: I fail to see the benefit. The NW method is too simple to be given its own task page, so it's nice to have it included in this one. More sophisticated methods can be handled in separate tasks and we can link to them from here. [[User:Fwend|Fwend]] ([[User talk:Fwend|talk]]) 12:47, 18 September 2014 (UTC)
:: Also I think it's better for testing purposes if the initial feasible solution is off the mark. That way your optimal solution routine is more pressured to produce a correct result than when you have an estimate that may already be correct to begin with. [[User:Fwend|Fwend]] ([[User talk:Fwend|talk]]) 13:16, 18 September 2014 (UTC)
::: This task could have used examples which better exercised the concepts. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 14:40, 4 February 2023 (UTC)
==Regarding The Java/D Versions==
| |||