Hi,
I have another problem for you, which is a case of the General
Assignment Problem (GAP):
http://0402.netclime.net/1_5/Z/P/5/AssignementProblem_g.pdf
Thanks.

---

Clarification of Question by ghassane-ga on 22 Oct 2004 18:43 PDT

Hi MathTalk,
i made a mistake in the file.
We have to found the Matrix Lamnda in function for matrix X=[x_{i,j}]
which miximize the objective function.
sorry :).

regards, ghassane-ga.

---

Clarification of Question by ghassane-ga on 22 Oct 2004 18:47 PDT

Hi,
I correct the problem in the file:
http://0302.netclime.net/1_5/150427/1098489625_20078/AssignementProblem_g.pdf
If the link changes, i'll give you the new location, thanks.

regards, ghassane-ga.

---

Clarification of Question by ghassane-ga on 22 Oct 2004 21:43 PDT

Hi,
sorry for my many clarifications, but this is the true link to problem's file:
http://0301.netclime.net/1_5/J/T/B/AssignementProblem_g.pdf
thanks.

regards, ghassane-ga.

---

Request for Question Clarification by mathtalk-ga on 23 Oct 2004 16:53 PDT

Hi, ghassane-ga:

I've looked over the problem definition.  What exactly are you asking
about it?  There is no difficulty in formulating a search strategy to
determine the optimal solution with complexity O(N^M), which in
practice can benefit strongly from a combination of a "greedy"
algorithm and pruning.

regards, mathtalk-ga

---

Clarification of Question by ghassane-ga on 23 Oct 2004 19:07 PDT

Hi,
I'm asking if is it possible to found an analytic solution (lambda in
function of X).
If not, so forgot this problem and focus on the other problem which is
more important for me :).

regards, ghassane-ga.

---

Clarification of Question by ghassane-ga on 24 Oct 2004 09:32 PDT

Hi,
ok i see. So, there is no analytic solution.
thank you.
I hope that you will find solution for the first problem :).
thanks.

regards, ghassane-ga.

Even the simplest possible case (1 by 2) would not make lambda
"analytic" in X, as the "function" must discontinously jump from
picking one entry to the other based on which is largest.

regards, mathtalk-ga