Get Time
Search | Watch Thread  |  My Post History  |  My Watches  |  User Settings
View: Flat (newest first)  | Threaded  | Tree
Previous Thread  |  Next Thread
about labeling function | Reply
Hi people, i only have a question about the algorithm...

The thing that is not clear to me is, how the labeling function works?

I mean, why this function always make the algorithm get the matching with the biggest weight, after the updates all the labels?

Is there a proof that explain this labeling function?, i mean does somebody have some article that proof this?

Sorry for my english
Re: about labeling function (response to post by WRBH) | Reply
I stuck at the same aspect. I tought the whole weekend about that and, by keeping the worker/job-example, I totally do not understand how the labeling works.

In all other examples the labels are just numbers itself but when I number the vertixes in a straigt series the condition "l(x)+l(y)=> w(x,y)" respectively "lx[x] = max(lx[x], cost[x][y])" makes no sense ?!

Can anoybody please help ?
Re: about labeling function (response to post by SkullCrossbone) | Reply
Ok, I guess I got it how it is meant in the explanation :)

But a further question: The labeling is obviously only done for one job as follows: Take the max weight of each worker to one job and call this labeling.

But what is with the other jobs ?? How does the code know the weight of the edges of the other jobs ?