Rosenbrock Function

Not all unimodel functions are simple to optimize. The fitness landscape is simple from afar but banana shaped when close to minimum. Rosenbrock's variables are strongly dependent and gradient information often misleads algorithms.

It is unimodal function
Minima : [0, 0, ..., 0]
Maxima : 2508237869
Maximum Dimensionality : 30

Function definition

3D Plot

    clc; clear
    x = -10:10; y = -10:10;
    xy=zeros(length(x)*length(y),2);
    cntpts=0;
    for ii=1:length(x)
        for jj=1:length(y)
            cntpts=cntpts+1;
            xy(cntpts,1:2)=[x(ii),y(jj)];
        end
    end
    funval = Rosenbrock([xy]);
    funval = reshape(funval, length(y),length(x));
    surfc(x, y, funval)

Other Useful Functions

DeJong Rosenbrock Ackley Alpine Rastrigin Griewank Schaffer