I have two square matrices (LS,RS) in the form of generalized eigen-value problem as below:

LS*z=omega*RS*z

One time, I am using inverse of RS multiplying by LS to get the eigen values as

MS := MatrixInverse(RS) . LS:

VL1, VR1 := Eigenvectors(MS):

Next time, I am using direct method as below to get the eigen values

VL2, VR2 := Eigenvectors(LS, RS);

I am wondering there are meaningful differences between VL1 and VL2 as well as VR1 and VR2.

Does anybody know why?