Section 6.9.2 of the CCX 2.17 documentation mentions that when performing a *FREQUENCY analysis:
instead of looking for the smallest eigenfrequencies of the generalized eigenvalue problem, the largest eigenvalues of the inverse problem are determined. For large problems this results in execution times cut by about a factor of 100 (!)
Could someone please explain (or point to a reference that explains) how this inverse problem is formulated? How are the global mass and stiffness matrices manipulated and the generalized eigenvalue equation formed for this inverse problem? In other words, could someone please use equations to explain to me how this works (or point to a reference that explains it)?
Many thanks for any help you can provide!