This is the power iteration method to find the maximum eigenvalue/eigenvector a n-by-n matrix. This method doesn't require the matrix to be Hermitian for the maximum eigenvalue/eigenvecor.
But it DOES require the matrix to be Hermitian for the minimum eigenvalue/vector. This approximation method may be improved by setting a tolerance (currently the iteration is controlled by the number of iterations, MAX).
Example: c = [1 0.5 0.2;0.5 1 0.5; 0.2 0.5 1];
then [u,v] = mPowerEig(c,0) is to find the largest eigenvalue/vector
and [u,v] = mPowerEig(c,1) is to find the minimum eigenvalue/vector
Reference: G.H. Golub, C.F. Van Load, "Matrix Computation"