Thanks for answering, but I tried in Maple 11 increasing depthlimit, but it still doesn't work. Could it be something from my computer? Because I use Maple 11 here, and Maple 10 in the university, and there the comand works, here it doesn't.

(I copied the comand on Maple, and I'm putting here what happened when I tried)

> with(Optimization);

Maximize(40*a+60*b, {40*a+60*b <= 150}, assume = nonnegint, depthlimit = 10);

Warning, problem appears to be unbounded
                             [0, [a = 0, b = 0]]
 

Mario

I am unable to reproduce the error you are getting. Instead, I get the expected solution by increasing depthlimit.

However, I wouldn't assume right away that the problem is with your computer, although using a different system can have an effect. Most of the algorithms in the Optimization package are sensitive to floating-point errors and to the choice of initial point. That is just a fact of life with many numerical local optimization solvers. Occasionally, you'll get an unexpected error with these routines but more frequently, you might get a local minimum different from what you expect.

For this particular example, try using a different initial point and see if that helps to avoid the error. Indeed, with routines in the Optimization package, one should always provide an initial point if at all possible. You can also set infolevel[Optimization] to a value of 3 or higher to get some information about what is happening.

Paulina Chin
Maplesoft

This problem has only 8 feasible points: (0,0), (0,1), (0,2), (1,0), (1,1), (2,0), (2,1), (3,0) - in such cases the simplest solution would be to evaluate the function in all the feasible points, and then choose the maximum. It is strange that such a simple thing was not implemented.

Alec

I tried every thing you said, but it still doesn't work. So I will submit it to Maplesoft Support.

Thanks for all your help.

Mario