Nonlinear Optimization in Matlab

There are various matlab functions available for nonlinear optimizations as well. These can be utilized in more exotic FIR filter designs, such as designs driven more by perceptual criteria:

• The fsolve function in Octave, or the Matlab Optimization Toolbox, attempts to solve unconstrained, overdetermined, nonlinear systems of equations.
• The Octave function sqp handles constrained nonlinear optimization.
• The Octave optim package includes many additional functions such as leasqr for performing Levenberg-Marquardt nonlinear regression. (Say, e.g., which leasqr and explore its directory.)
• The Matlab Optimization Toolbox similarly contains many functions for optimization.

Again, this is only a sampling of what is available.