fmincon optimization with min/max criterion - constraint violation

1 view (last 30 days)
Hi all,
We are trying to implement the min/max criterion[1] in a static optimization problem using fmincon. The purpose is to compare different objective functions in how they solve a load sharing problem. But we face some 'strange' behaviour of the optimizer.
Where the more common objective functions remain in their bounds even if no feasible solution is found, the min/max criterion violates the constraints and the lower and upper bounds given. Up to the point where the solution violates the constraints the outcome seems reasonable.
Is it common that infeasible solutions violate the bounds set? And can this boundary violation (and constraint violation) have anything to do with the fact that the gradient and hessian of the objective function are zero?
[1] Rasmussen J. and Damsgaard M. and Voigt M., Muscle recruitment by the min/max criterion - a comparative numerical study, 2001, Journal of Biomechanics 34, 409-415
Best regards, Priti and Erik

Accepted Answer

Alan Weiss
Alan Weiss on 31 May 2013
I don't really understand what you are doing, but most fmincon algorithms satisfy bounds at every iteration. See the details here.
Alan Weiss
MATLAB mathematical toolbox documentation

More Answers (0)

Products

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!