1: \begin{abstract}

2: Three refined and refined harmonic extraction-based

3: Jacobi--Davidson (JD) type methods are proposed, and their

4: thick-restart algorithms with deflation and purgation are

5: developed to compute several generalized singular value

6: decomposition (GSVD) components of a large regular matrix pair.

7: The new methods are called refined cross product-free (RCPF),

8: refined cross product-free harmonic (RCPF-harmonic) and refined

9: inverse-free harmonic (RIF-harmonic) JDGSVD algorithms,

10: abbreviated as RCPF-JDGSVD, RCPF-HJDGSVD and RIF-HJDGSVD, respectively.

11: The new JDGSVD methods are more efficient than the corresponding

12: standard and harmonic extraction-based JDSVD methods proposed previously

13: by the authors, and can overcome the erratic behavior

14: and intrinsic possible non-convergence of the latter ones.

15: Numerical experiments illustrate

16: that RCPF-JDGSVD performs better for the computation of

17: extreme GSVD components while RCPF-HJDGSVD and RIF-HJDGSVD suit better for

18: that of interior GSVD components.

19: \end{abstract}

20: