| C. Zhu, R.H. Byrd, P. Lu and J. Nocedal, "LBFGS-B: Fortran subroutines for large-scale bound constrained optimization", Report NAM-11, EECS Department, Northwestern University, 1994. 24 |
....c i (l i ) s do not depend on the iterations. However, these c i (l i ) s are much larger than the optimal curvatures, so more iterations are required for PS,M,CD than PS,O,CD to converge. We also compared the PSCD algorithms to the general purpose constrained Quasi Newton algorithm (LBFGS) [11], and the functional substitution coordinate descent (FSCD) 12, 13] algorithm in Figures 1 and 2. Although the LBFGS algorithm takes about 25 less CPU time (0.88 seconds) per iteration than PSCD algorithms, it did not converge as fast as the proposed algorithms. FSCD algorithm (which is ....
C. Zhu, R. H. Byrd, P. Lu, and J. Nocedal, "L-BFGS-B: Fortran subroutine for large-scale bound contrained optimization, " Technical Report ?, Dept. EECS, Northwestern Univ., July 1996.
....c i (l n i ) s do not depend on the iterations. However, these c i (l n i ) s are much larger than the optimal curvatures, so more iterations are required for PS,M,CD than PS,O,CD to converge. We also compared the PSCD algorithms to the general purpose constrained Quasi Newton algorithm (LBFGS) [11], and the functional substitution coordinate descent (FSCD) 12, 13] algorithm in Figures 1 and 2. Although the LBFGS algorithm takes about 25 less CPU time (0.88 seconds) per iteration than PSCD algorithms, it did not converge as fast as the proposed algorithms. FSCD algorithm (which is ....
C. Zhu, R. H. Byrd, P. Lu, and J. Nocedal, "L-BFGS-B: Fortran subroutine for large-scale bound contrained optimization, " Technical Report ?, Dept. EECS, Northwestern Univ., July 1996.
....is available, or when significant advantage can be taken of structure. However this study appears to indicate that, in terms of computing time, the new limited memory code is very competitive with all versions of LANCELOT on many problems. A code implementing the new algorithm is described in [23], and can be obtained by contacting the authors at nocedal eecs.nwu.edu. 22 7. ....
C. Zhu, R.H. Byrd, P. Lu and J. Nocedal, "LBFGS-B: Fortran subroutines for large-scale bound constrained optimization", Report NAM-11, EECS Department, Northwestern University, 1994. 24
....is available, or when significant advantage can be taken of structure. However this study appears to indicate that, in terms of computing time, the new limited memory code is very competitive with all versions of LANCELOT on many problems. A code implementing the new algorithm is described in [23], and can be obtained by contacting the authors at nocedal eecs.nwu.edu. 7. ....
C. Zhu, R.H. Byrd, P. Lu and J. Nocedal, "LBFGS-B: Fortran subroutines for large-scale bound constrained optimization", Report NAM-11, EECS Department, Northwestern University, 1994.
No context found.
C. Zhu, R. H. Byrd, P. Lu, and J. Nocedal, "LBFGS-B: Fortran subroutines for large-scale bound constrained optimization," Tech. Rep. NAM-11, EECS Department, Northwestern University, 1994.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC