Math Forum / Mathematics / Operations Research / November 2008

I am trying to optimize a model M. And when I use root mean square
error(RMSE) as the cost function, I always obtain a better model than
when I use  mean absolute error as the cost function. I am wondering
why it is that.

I'm working on a project that uses a Mixed Integer Programming model
for many good reasons. However, one of the requirements is to place
sets of multiple colors in vertices of an undirected graph (one piece
per vertex), so that they form connected components.

I posted this on sci.stat.math as well, but did not get any response.
Consider the following problem:
X is a random variable following Erlang distribution with shape
parameter M and rate parameter lambda.

For the sake of saving some expense,
we are thinking using the open source for linear programming such as
COIN-OR, GNU, and CPL.
We built this s/w with C#, so if the open source works with C#, that

I'm an advanced user of ILOG OPL studio, which I find very powerfull
to model complex data structures, including sparse ones. I put a
screeshot sample on my web site http://www.rostudel.com/ (screenshot
then "OPL and sparsity") that creates variables on a set of people

I have a simplified question looked like the following and
hope someone can point me where I can start.
I like to know how to model this problem and then solve it.
    |    |    |    |

Would you please help me to solve my problem
I need to compute a very good lower bound of a MIP with cplex, I'm not
interrested to the optimal solution
Is it possible?

I have a question about penalty function method. In the convergence
theorem of penalty function, it says, any limit point of a sequence
generated by the penalty method is also an optimal solution of
original nonlinear programming problem.

I am trying the evaluation version of TOMLAB to see if it fulfills the
requirements of our research group. I have a wireless network
modelling problem defined in lp format of lp_solve (http://
lpsolve.sourceforge.net/5.5/lp-format.htm)

Can anyone recommend a website with some easy Operations Research (or
other optimisation) puzzles which can readily be tackled with lpsolve
just for fun, please?
Thanks for any suggestions!

This problem is very very important to me! I realy need help!
any one can help me with the following problem?
it is a nonlinear programming problem
====================================

Does anyone know if there is any modules in R other than lpSolve.
I specifically want to look at the knapsack problem with multiple
constraints and boundary conditions
So far I haven't come across anything, so any advice welcome

I don't know if my question is op-research or not - it seems like it
could be, but I don't know that much about op-research. I asked in
sci.stat.math but got no response. So here goes:
It would take some time to explain my real application, so I'll use

Hello, there.
We have around 20 c# files that are used for creating Objective
functions. It is just that each file is used for a term in the
objective function. Each class looks like:

I am tailoring a somewhat complex MILP transportation model which
involves
penalties for earliness and tardiness and when putting in the
penalties I have