*********************************************************************** * * * PLIV - A LIMITED MEMORY VARIABLE METRIC ALGORITHM WITH * * MULTIPLE CORRECTIONS FOR LARGE-SCALE OPTIMIZATION. * * * *********************************************************************** 1. Introduction: ---------------- The double-precision FORTRAN 77 basic subroutine PLIV is designed to find a close approximation to a local minimum of a nonlinear function F(X) with simple bounds on variables. Here X is a vector of NF variables and F(X) is a smooth function. We suppose that NF is large but the sparsity pattern of the Hessian matrix is not known (or the Hessian matrix is dense). Simple bounds are assumed in the form X(I) unbounded if IX(I) = 0, XL(I) <= X(I) if IX(I) = 1, X(I) <= XU(I) if IX(I) = 2, XL(I) <= X(I) <= XU(I) if IX(I) = 3, XL(I) = X(I) = XU(I) if IX(I) = 5, where 1 <= I <= NF. To simplify user's work, two additional easy to use subroutines are added. They call the basic general subroutine PLIV: PLIVU - unconstrained large-scale optimization, PLIVS - large-scale optimization with simple bounds. All subroutines contain a description of formal parameters and extensive comments. Furthermore, two test programs TLIVU and TLIVS are included, which contain several test problems (see e.g. [2]). These test programs serve as examples for using the subroutines, verify their correctness and demonstrate their efficiency. In this short guide, we describe all subroutines which can be called from the user's program. A detailed description of the method is given in [1]. In the description of formal parameters, we introduce a type of the argument that specifies whether the argument must have a value defined on entry to the subroutine (I), whether it is a value which will be returned (O), or both (U), or whether it is an auxiliary value (A). Besides formal parameters, we can use a COMMON /STAT/ block containing statistical information. This block, used in each subroutine has the following form: COMMON /STAT/ NRES,NDEC,NIN,NIT,NFV,NFG,NFH The arguments have the following meaning: Argument Type Significance ---------------------------------------------------------------------- NRES O Positive INTEGER variable that indicates the number of restarts. NDEC O Positive INTEGER variable that indicates the number of matrix decompositions. NIN O Positive INTEGER variable that indicates the number of inner iterations (for solving linear systems). NIT O Positive INTEGER variable that indicates the number of iterations. NFV O Positive INTEGER variable that indicates the number of function evaluations. NFG O Positive INTEGER variable that indicates the number of gradient evaluations. NFH O Positive INTEGER variable that indicates the number of Hessian evaluations. 2. Subroutines PLIVU, PLIVS: ---------------------------- The calling sequences are CALL PLIVU(NF,X,IPAR,RPAR,F,GMAX,IPRNT,ITERM) CALL PLIVS(NF,X,IX,XL,XU,IPAR,RPAR,F,GMAX,IPRNT,ITERM) The arguments have the following meaning. Argument Type Significance ---------------------------------------------------------------------- NF I Positive INTEGER variable that specifies the number of variables of the objective function. X(NF) U On input, DOUBLE PRECISION vector with the initial estimate to the solution. On output, the approximation to the minimum. IX(NF) I On input (significant only for PLIVS) INTEGER vector containing the simple bounds types: IX(I)=0 - the variable X(I) is unbounded, IX(I)=1 - the lower bound X(I) >= XL(I), IX(I)=2 - the upper bound X(I) <= XU(I), IX(I)=3 - the two side bound XL(I) <= X(I) <= XU(I), IX(I)=5 - the variable X(I) is fixed (given by its initial estimate). XL(NF) I DOUBLE PRECISION vector with lower bounds for variables (significant only for PLIVS). XU(NF) I DOUBLE PRECISION vector with upper bounds for variables (significant only for PLIVS). IPAR(7) U INTEGER parameters: IPAR(1)=MIT, IPAR(2)=MFV, IPAR(3)-unused, IPAR(4)=IEST, IPAR(5)-MET, IPAR(6)-unused, IPAR(7)=MF. Parameters MIT, MFV, IEST, MF are described in Section 3 together with other parameters of the subroutine PLIV. RPAR(9) U DOUBLE PRECISION parameters: RPAR(1)=XMAX, RPAR(2)=TOLX, RPAR(3)=TOLF, RPAR(4)=TOLB, RPAR(5)=TOLG, RPAR(6)=FMIN, RPAR(7)-unused, RPAR(6)-unused, RPAR(9)-unused. Parameters XMAX, TOLX, TOLF, TOLB, TOLG, FMIN are described in Section 3 together with other parameters of the subroutine PLIV. F O DOUBLE PRECISION value of the objective function at the solution X. GMAX O DOUBLE PRECISION maximum absolute value of a partial derivative of the objective function. IPRNT I INTEGER variable that specifies PRINT: IPRNT= 0 - print is suppressed, IPRNT= 1 - basic print of final results, IPRNT=-1 - extended print of final results, IPRNT= 2 - basic print of intermediate and final results, IPRNT=-2 - extended print of intermediate and final results. ITERM O INTEGER variable that indicates the cause of termination: ITERM= 1 - if |X - XO| was less than or equal to TOLX in two subsequent iterations, ITERM= 2 - if |F - FO| was less than or equal to TOLF in two subsequent iterations, ITERM= 3 - if F is less than or equal to TOLB, ITERM= 4 - if GMAX is less than or equal to TOLG, ITERM= 6 - if termination criterion was not satisfied, but the solution is probably acceptable, ITERM=11 - if NIT exceeded MIT, ITERM=12 - if NFV exceeded MFV, ITERM< 0 - if the method failed. The subroutines PLIVU, PLIVS require the user supplied subroutines OBJ and DOBJ that define the objective function and its gradient and have the form SUBROUTINE OBJ(NF,X,F) SUBROUTINE DOBJ(NF,X,G) The arguments of the user supplied subroutines have the following meaning. Argument Type Significance ---------------------------------------------------------------------- NF I Positive INTEGER variable that specifies the number of variables of the objective function. X(NF) I DOUBLE PRECISION an estimate to the solution. F O DOUBLE PRECISION value of the objective function at the point X. G(NF) O DOUBLE PRECISION gradient of the objective function at the point X. 3. Subroutine PLIV: ------------------- This general subroutine is called from all subroutines described in Section 2. The calling sequence is CALL PLIV(NF,NB,X,IX,XL,XU,GF,S,XO,GO,XM,GM,XR,GR,IW,RW,XMAX,TOLX, & TOLF,TOLB,TOLG,FMIN,GMAX,F,MIT,MFV,IEST,MET,MF,IPRNT,ITERM) The arguments NF, NB, X, IX, XL, XU, GMAX, F, IPRNT, ITERM, have the same meaning as in Section 2. Other arguments have the following meaning: Argument Type Significance ---------------------------------------------------------------------- GF(NF) A DOUBLE PRECISION gradient of the objective function. S(NF) A DOUBLE PRECISION direction vector. XO(NF) A DOUBLE PRECISION array which contains increments of variables. GO(NF) A DOUBLE PRECISION array which contains increments of gradients. XM(NF*MF) A DOUBLE PRECISION array which contains previous increments of variables. GM(NF*MF) A DOUBLE PRECISION array which contains previous increments of gradients. XR(NZR) A DOUBLE PRECISION Auxiliary array (NCR is equal to 2*NF*NF). GR(NGR) A DOUBLE PRECISION Auxiliary array (NCR is equal to NF*(NF+1)/2). IW(MF*MF) A INTEGER Auxiliary array. RW(9*MF) A DOUBLE PRECISION Auxiliary array. XMAX U DOUBLE PRECISION maximum stepsize; the choice XMAX=0 causes that the default value 1.0D+16 will be taken. TOLX U DOUBLE PRECISION tolerance for the change of the coordinate vector X; the choice TOLX=0 causes that the default value TOLX=1.0D-16 will be taken. TOLF U DOUBLE PRECISION tolerance for the change of function values; the choice TOLF=0 causes that the default value TOLF=1.0D-14 will be taken. TOLB U DOUBLE PRECISION minimum acceptable function value; the choice TOLB=0 causes that the default value TOLB=FMIN+1.0D-16 will be taken. TOLG U DOUBLE PRECISION tolerance for the Lagrangian function gradient; the choice TOLG=0 causes that the default value TOLG=1.0D-6 will be taken. FMIN U DOUBLE PRECISION lower bound for the minimum function value. It is significant only if IEST=1. If IEST=0, the default value FMIN=-1.0D+60 will be taken. MIT U INTEGER variable that specifies the maximum number of iterations; the choice MIT=0 causes that the default value 9000 will be taken. MFV U INTEGER variable that specifies the maximum number of function evaluations; the choice MFV=0 causes that the default value 9000 will be taken. IEST I INTEGER estimation of the minimum functiom value for the line search: IEST=0 - estimation is not used, IEST=1 - lower bound FMIN is used as an estimation for the minimum function value. MET I Maximum number of correction steps; the choice TOLX=0 causes that the default value MET=4 will be taken. MF U The number of limited-memory variable metric updates in each iteration (they use 2*MF stored vectors). The choice MF=0 causes that the default value MF=5 will be taken. The choice of parameter XMAX can be sensitive in many cases. First, the objective function can be evaluated only in a relatively small region (if it contains exponentials) so that the maximum stepsize is necessary. Secondly, the problem can be very ill-conditioned far from the solution point so that large steps can be unsuitable. Finally, if the problem has more local solutions, a suitably chosen maximum stepsize can lead to obtaining a better local solution. The subroutine PLIV requires the user supplied subroutines OBJ and DOBJ which are described in Section 2. 4. Verification of the subroutines: ----------------------------------- Subroutine PLIVU can be verified and tested using the program TLIVU. This program calls the subroutines TIUD14 (initiation), TFFU14 (function evaluation) and TFGU14 (gradient evaluation) containing 22 unconstrained test problems with at most 1000 variables [2]. The results obtained by the program TLIVU on a PC computer with Microsoft Power Station Fortran compiler have the following form. NIT= 4957 NFV= 5479 NFG= 5479 F= 0.731101738E-13 G= 0.882E-06 ITERM= 4 NIT= 353 NFV= 411 NFG= 411 F= 14.9944763 G= 0.101E-04 ITERM= 2 NIT= 113 NFV= 122 NFG= 122 F= 0.575809315E-12 G= 0.119E-06 ITERM= 4 NIT= 107 NFV= 112 NFG= 112 F= 269.499543 G= 0.284E-05 ITERM= 2 NIT= 23 NFV= 26 NFG= 26 F= 0.457722424E-12 G= 0.268E-06 ITERM= 4 NIT= 30 NFV= 31 NFG= 31 F= 0.196012198E-10 G= 0.929E-06 ITERM= 4 NIT= 39 NFV= 44 NFG= 44 F= 335.137433 G= 0.144E-05 ITERM= 2 NIT= 28 NFV= 31 NFG= 31 F= 761774.954 G= 0.164E-03 ITERM= 2 NIT= 13 NFV= 16 NFG= 16 F= 316.436141 G= 0.291E-06 ITERM= 4 NIT= 2007 NFV= 2029 NFG= 2029 F= -124.750000 G= 0.675E-05 ITERM= 2 NIT= 103 NFV= 124 NFG= 124 F= 10.7765879 G= 0.580E-06 ITERM= 4 NIT= 242 NFV= 265 NFG= 265 F= 982.273617 G= 0.170E-04 ITERM= 2 NIT= 7 NFV= 8 NFG= 8 F= 0.279448694E-15 G= 0.311E-07 ITERM= 4 NIT= 7 NFV= 9 NFG= 9 F= 0.128843969E-08 G= 0.991E-06 ITERM= 4 NIT= 996 NFV= 1005 NFG= 1005 F= 1.92401599 G= 0.731E-06 ITERM= 4 NIT= 199 NFV= 206 NFG= 206 F= -427.404476 G= 0.511E-05 ITERM= 2 NIT= 503 NFV= 505 NFG= 505 F=-0.379921091E-01 G= 0.929E-06 ITERM= 4 NIT= 501 NFV= 504 NFG= 504 F=-0.245741193E-01 G= 0.117E-13 ITERM= 4 NIT= 501 NFV= 504 NFG= 504 F= 59.5986241 G= 0.661E-10 ITERM= 4 NIT= 2178 NFV= 2182 NFG= 2182 F= -1.00013520 G= 0.863E-06 ITERM= 4 NIT= 2231 NFV= 2234 NFG= 2234 F= 2.13866377 G= 0.905E-06 ITERM= 4 NIT= 1369 NFV= 1399 NFG= 1399 F= 1.00000000 G= 0.938E-06 ITERM= 4 NITER =16507 NFVAL =17246 NGVAL =17246 NSUCC = 22 TIME= 0:00:03.28 The rows corresponding to individual test problems contain the number of iterations NIT, the number of function evaluations NFV, the number of gradient evaluations NFG, the final value of the objective function F, the norm of gradient G and the cause of termination ITERM. Subroutine PLIVS can be verified and tested using the program TLIVS. This program calls the subroutines TIUD14 (initiation), TFFU14 (function evaluation), TFGU14 (gradient evaluation) containing 22 box constrained test problems with at most 1000 variables [2]. The results obtained by the program TLIVS on a PC computer with Microsoft Power Station Fortran compiler have the following form. NIT= 5112 NFV= 5627 NFG= 5627 F= 0.00000000 G= 0.000E+00 ITERM= 3 NIT= 2207 NFV= 2853 NFG= 2853 F= 3926.45961 G= 0.183E-04 ITERM= 2 NIT= 113 NFV= 125 NFG= 125 F= 0.312399241E-12 G= 0.299E-06 ITERM= 4 NIT= 59 NFV= 72 NFG= 72 F= 269.522686 G= 0.742E-06 ITERM= 4 NIT= 23 NFV= 26 NFG= 26 F= 0.457722424E-12 G= 0.268E-06 ITERM= 4 NIT= 30 NFV= 31 NFG= 31 F= 0.196012198E-10 G= 0.929E-06 ITERM= 4 NIT= 40 NFV= 45 NFG= 45 F= 336.134919 G= 0.380E-06 ITERM= 4 NIT= 54 NFV= 56 NFG= 56 F= 761925.725 G= 0.227E-03 ITERM= 2 NIT= 505 NFV= 507 NFG= 507 F= 428.056916 G= 0.272E-06 ITERM= 4 NIT= 1014 NFV= 1039 NFG= 1039 F= -81.0572252 G= 0.672E-05 ITERM= 2 NIT= 13 NFV= 23 NFG= 23 F= 96517.2947 G= 0.971E-06 ITERM= 4 NIT= 66 NFV= 81 NFG= 81 F= 4994.21410 G= 0.712E-06 ITERM= 4 NIT= 7 NFV= 8 NFG= 8 F= 0.279448694E-15 G= 0.311E-07 ITERM= 4 NIT= 7 NFV= 9 NFG= 9 F= 0.128843969E-08 G= 0.991E-06 ITERM= 4 NIT= 996 NFV= 1005 NFG= 1005 F= 1.92401599 G= 0.731E-06 ITERM= 4 NIT= 180 NFV= 181 NFG= 181 F= -427.391653 G= 0.108E-04 ITERM= 2 NIT= 503 NFV= 505 NFG= 505 F=-0.379921091E-01 G= 0.929E-06 ITERM= 4 NIT= 501 NFV= 504 NFG= 504 F=-0.245741193E-01 G= 0.117E-13 ITERM= 4 NIT= 915 NFV= 920 NFG= 920 F= 1654.94525 G= 0.224E-04 ITERM= 2 NIT= 2022 NFV= 2025 NFG= 2025 F= -1.00013520 G= 0.973E-06 ITERM= 4 NIT= 1385 NFV= 1386 NFG= 1386 F= 2.41354873 G= 0.991E-06 ITERM= 4 NIT= 1488 NFV= 1530 NFG= 1530 F= 1.00000000 G= 0.986E-06 ITERM= 4 NITER =17240 NFVAL =18558 NGVAL =18558 NSUCC = 22 TIME= 0:00:03.55 References: ----------- [1] Vlcek J., Luksan L.: A modified limited-memory BNS method for unconstrained minimization based on the conjugate directions idea. Techical Report V-1203, Prague, ICS AS CR, 2014. [2] Luksan L., Vlcek J.: Sparse and partially separable test problems for unconstrained and equality constrained optimization. Research Report V-767, Institute of Computer Science, Academy of Sciences of the Czech Republic, Prague, Czech Republic, 1998.