MATLAB Examples

# Generalized eigenvalue problem

Example 4.2 from Numerically solving polynomial systems with Bertini, by Daniel J. Bates, Jonathan D. Haunstein, Andrew J. Sommese and Charles W. Wampler (SIAM 2013).

Solve the generalized eigenvalue problem

$\mu A v = \lambda B v.$

A singular problem is chosen in which the solutions are $(\mu,\lambda) = (0,1)$ and $(\mu,\lambda) =(0,1)$, so the problem cannot be reduced to having a single eigenvalue. This problem is homogeneous in two separate sets of coordinates. Note that, in the definition of poly_system two groups of variables are assigned to hom_variable_group: (mu, lambda) and v.

polysyms mu0 lambda; % Use 'mu0' because 'mu' is a Matlab toolbox command v = polysym('v',[1 2]); fname = polysym('f',[1 2]); A = [1 2; 2 4]; B = [4 -2; -2 1]; fval = mu0*A*v.' - lambda*B*v.'; poly_system = BertiniLab('function_def',fname(:),'function_def',fval(:), ... 'hom_variable_group',{[mu0 lambda],v}); poly_system = solve(poly_system); sols = poly_system.match_solutions('raw_solutions',mu0,lambda,v); 

This problem has eigenvalue pairs mu0 and lambda:

musols = double(sols.mu0); lambdasols = double(sols.lambda); fprintf('%14s %16s\n','mu','lambda') fprintf('%15.4f %15.4f\n',[real(musols) real(lambdasols)].') 
 mu lambda 1.0000 -0.0000 0.0000 1.0000