diff --git a/CHNRSNBM.SIF b/CHNRSNBM.SIF new file mode 100644 index 0000000..b051d75 --- /dev/null +++ b/CHNRSNBM.SIF @@ -0,0 +1,142 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME CHNRSNBM + +* Problem : +* -------- +* A variable dimension version of the chained Rosenbrock function (CHNROSNB) +* by Luksan et al. + +* Source: problem 27 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* that is an extension of that proposed in +* Ph.L. Toint, +* "Some numerical results using a sparse matrix updating formula in +* unconstrained optimization", +* Mathematics of Computation, vol. 32(114), pp. 839-852, 1978. + +* See also Buckley#46 (n = 25) (p. 45). +* SIF input: Ph. Toint, Dec 1989. +* this version Nick Gould, June, 2013 + +* classification SUR2-AN-V-0 + +* Number of variables ( at most 50) + +*IE N 10 $-PARAMETER original value +*IE N 25 $-PARAMETER + IE N 50 $-PARAMETER + +* other parameter definitions + + IE 1 1 + IE 2 2 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + DO I 2 N + IA I-1 I -1 + XN SQ(I) X(I-1) 1.0 + RI RI I + R( SINI SIN RI + RA ALPHA SINI 1.5 + R* AI2 ALPHA ALPHA + RM 16AI2 AI2 16.0 + RD SCL 16AI2 1.0 + ZN SQ(I) 'SCALE' SCL + XN B(I) X(I) 1.0 + ND + +CONSTANTS + + DO I 2 N + X CHNRSNBM B(I) 1.0 + ND + +BOUNDS + + FR CHNRSNBM 'DEFAULT' + +START POINT + + XV CHNROSMB 'DEFAULT' -1.0 + +ELEMENT TYPE + + EV ETYPE V1 + +ELEMENT USES + + XT 'DEFAULT' ETYPE + + DO I 2 N + ZV ELA(I) V1 X(I) + ND + +GROUP TYPE + + GV L2 GVAR + +GROUP USES + + XT 'DEFAULT' L2 + + DO I 2 N + XE SQ(I) ELA(I) + ND + +OBJECT BOUND + + LO CHNRSNBM 0.0 + +* Solution + +*LO SOLTN 0.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS CHNRSNBM + +INDIVIDUALS + + T ETYPE + F - V1 ** 2 + G V1 - 2.0 * V1 + H V1 V1 - 2.0 + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS CHNRSNBM + +INDIVIDUALS + + T L2 + + F GVAR * GVAR + G GVAR + GVAR + H 2.0 + +ENDATA diff --git a/DIXMAANM.SIF b/DIXMAANM.SIF new file mode 100644 index 0000000..372f8ee --- /dev/null +++ b/DIXMAANM.SIF @@ -0,0 +1,255 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME DIXMAANM + +* Problem : +* ********* +* A variant on the Dixon-Maany test problem (version I) + +* Source: +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* (problem 19) based on + +* L.C.W. Dixon and Z. Maany, +* "A family of test problems with sparse Hessians for unconstrained +* optimization", +* TR 206, Numerical Optimization Centre, Hatfield Polytechnic, 1988. + +* SIF input: Ph. Toint, Dec 1989. +* correction by Ph. Shott, January 1995. +* this version Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* M is equal to the third of the number of variables + +*IE M 5 $-PARAMETER n = 15 original value +*IE M 30 $-PARAMETER n = 90 +*IE M 100 $-PARAMETER n = 300 +*IE M 500 $-PARAMETER n = 1500 + IE M 1000 $-PARAMETER n = 3000 +*IE M 3000 $-PARAMETER n = 9000 + IE M 5 $-PARAMETER n = 15 original value + +* N is the number of variables + + IM N M 3 + +* Problem parameters + + RE ALPHA 1.0 + RE BETA 0.0 + RE GAMMA 0.125 + RE DELTA 0.125 + +* K-set 3 + + IE K1 2 + IE K2 1 + IE K3 1 + IE K4 2 + +* Other parameters + + RI RN N + IA N-1 N -1 + I+ 2M M M + + IE 1 1 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + N GA + N GB + N GC + N GD + +CONSTANTS + + DIXMAANM GA -1.0 + +BOUNDS + + FR DIXMAANM 'DEFAULT' + +START POINT + + XV DIXMAANM 'DEFAULT' 2.0 + +ELEMENT TYPE + + EV SQ X + EV SQB X Y + EV SQC X Y + EV 2PR X Y + +ELEMENT USES + +* First group + + DO I 1 N + XT A(I) SQ + ZV A(I) X X(I) + ND + +* Second group + + DO I 1 N-1 + IA I+1 I 1 + XT B(I) SQB + ZV B(I) X X(I) + ZV B(I) Y X(I+1) + ND + +* Third group + + DO I 1 2M + I+ I+M I M + XT C(I) SQC + ZV C(I) X X(I) + ZV C(I) Y X(I+M) + ND + +* Fourth group + + DO I 1 M + I+ I+2M I 2M + XT D(I) 2PR + ZV D(I) X X(I) + ZV D(I) Y X(I+2M) + ND + +GROUP USES + +* First group + + DO I 1 N + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K1 + R* TMP TMP I/N + OD J + R* AI TMP ALPHA + ZE GA A(I) AI + ND + +* Second group + + DO I 1 N-1 + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K2 + R* TMP TMP I/N + OD J + R* BI TMP BETA + ZE GB B(I) BI + ND + +* Third group + + DO I 1 2M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K3 + R* TMP TMP I/N + OD J + R* CI TMP GAMMA + ZE GC C(I) CI + ND + +* Fourth group + + DO I 1 M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K4 + R* TMP TMP I/N + OD J + R* DI TMP DELTA + ZE GD D(I) DI + ND + +OBJECT BOUND + + LO DIXMAANM 0.0 + +* Solution + +*LO SOLTN 1.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS DIXMAANM + +TEMPORARIES + + R F1 + R F2 + R DF2DY + +INDIVIDUALS + +* First type + + T SQ + F X * X + G X X + X + H X X 2.0 + +* Second type + + T SQB + A F1 X * X + A F2 Y + Y * Y + A DF2DY 1.0 + 2.0 * Y + F F1 * F2 * F2 + G X 2.0 * X * F2 * F2 + G Y 2.0 * F1 * F2 * DF2DY + H X X 2.0 * F2 * F2 + H X Y 4.0 * X * DF2DY * F2 + H Y Y 4.0 * F1 * F2 + + H+ 2.0 * F1 * DF2DY * DF2DY + +* Third type + + T SQC + A F1 X * X + A F2 Y**4 + F F1 * F2 + G X 2.0 * X * F2 + G Y 4.0 * F1 * Y**3 + H X X 2.0 * F2 + H X Y 8.0 * X * Y**3 + H Y Y 12.0 * F1 * Y**2 + +* Fourth type + + T 2PR + F X * Y + G X Y + G Y X + H X Y 1.0 + +ENDATA diff --git a/DIXMAANN.SIF b/DIXMAANN.SIF new file mode 100644 index 0000000..caf56b3 --- /dev/null +++ b/DIXMAANN.SIF @@ -0,0 +1,255 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME DIXMAANN + +* Problem : +* ********* +* A variant on the Dixon-Maany test problem (version I) + +* Source: +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* (problem 20) based on + +* L.C.W. Dixon and Z. Maany, +* "A family of test problems with sparse Hessians for unconstrained +* optimization", +* TR 206, Numerical Optimization Centre, Hatfield Polytechnic, 1988. + +* SIF input: Ph. Toint, Dec 1989. +* correction by Ph. Shott, January 1995. +* this version Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* M is equal to the third of the number of variables + +*IE M 5 $-PARAMETER n = 15 original value +*IE M 30 $-PARAMETER n = 90 +*IE M 100 $-PARAMETER n = 300 +*IE M 500 $-PARAMETER n = 1500 + IE M 1000 $-PARAMETER n = 3000 +*IE M 3000 $-PARAMETER n = 9000 + IE M 5 $-PARAMETER n = 15 original value + +* N is the number of variables + + IM N M 3 + +* Problem parameters + + RE ALPHA 1.0 + RE BETA 0.0625 + RE GAMMA 0.0625 + RE DELTA 0.0625 + +* K-set 3 + + IE K1 2 + IE K2 1 + IE K3 1 + IE K4 2 + +* Other parameters + + RI RN N + IA N-1 N -1 + I+ 2M M M + + IE 1 1 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + N GA + N GB + N GC + N GD + +CONSTANTS + + DIXMAANN GA -1.0 + +BOUNDS + + FR DIXMAANN 'DEFAULT' + +START POINT + + XV DIXMAANN 'DEFAULT' 2.0 + +ELEMENT TYPE + + EV SQ X + EV SQB X Y + EV SQC X Y + EV 2PR X Y + +ELEMENT USES + +* First group + + DO I 1 N + XT A(I) SQ + ZV A(I) X X(I) + ND + +* Second group + + DO I 1 N-1 + IA I+1 I 1 + XT B(I) SQB + ZV B(I) X X(I) + ZV B(I) Y X(I+1) + ND + +* Third group + + DO I 1 2M + I+ I+M I M + XT C(I) SQC + ZV C(I) X X(I) + ZV C(I) Y X(I+M) + ND + +* Fourth group + + DO I 1 M + I+ I+2M I 2M + XT D(I) 2PR + ZV D(I) X X(I) + ZV D(I) Y X(I+2M) + ND + +GROUP USES + +* First group + + DO I 1 N + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K1 + R* TMP TMP I/N + OD J + R* AI TMP ALPHA + ZE GA A(I) AI + ND + +* Second group + + DO I 1 N-1 + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K2 + R* TMP TMP I/N + OD J + R* BI TMP BETA + ZE GB B(I) BI + ND + +* Third group + + DO I 1 2M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K3 + R* TMP TMP I/N + OD J + R* CI TMP GAMMA + ZE GC C(I) CI + ND + +* Fourth group + + DO I 1 M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K4 + R* TMP TMP I/N + OD J + R* DI TMP DELTA + ZE GD D(I) DI + ND + +OBJECT BOUND + + LO DIXMAANN 0.0 + +* Solution + +*LO SOLTN 1.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS DIXMAANN + +TEMPORARIES + + R F1 + R F2 + R DF2DY + +INDIVIDUALS + +* First type + + T SQ + F X * X + G X X + X + H X X 2.0 + +* Second type + + T SQB + A F1 X * X + A F2 Y + Y * Y + A DF2DY 1.0 + 2.0 * Y + F F1 * F2 * F2 + G X 2.0 * X * F2 * F2 + G Y 2.0 * F1 * F2 * DF2DY + H X X 2.0 * F2 * F2 + H X Y 4.0 * X * DF2DY * F2 + H Y Y 4.0 * F1 * F2 + + H+ 2.0 * F1 * DF2DY * DF2DY + +* Third type + + T SQC + A F1 X * X + A F2 Y**4 + F F1 * F2 + G X 2.0 * X * F2 + G Y 4.0 * F1 * Y**3 + H X X 2.0 * F2 + H X Y 8.0 * X * Y**3 + H Y Y 12.0 * F1 * Y**2 + +* Fourth type + + T 2PR + F X * Y + G X Y + G Y X + H X Y 1.0 + +ENDATA diff --git a/DIXMAANO.SIF b/DIXMAANO.SIF new file mode 100644 index 0000000..c5965c9 --- /dev/null +++ b/DIXMAANO.SIF @@ -0,0 +1,255 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME DIXMAANO + +* Problem : +* ********* +* A variant on the Dixon-Maany test problem (version I) + +* Source: +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* (problem 21) based on + +* L.C.W. Dixon and Z. Maany, +* "A family of test problems with sparse Hessians for unconstrained +* optimization", +* TR 206, Numerical Optimization Centre, Hatfield Polytechnic, 1988. + +* SIF input: Ph. Toint, Dec 1989. +* correction by Ph. Shott, January 1995. +* this version Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* M is equal to the third of the number of variables + +*IE M 5 $-PARAMETER n = 15 original value +*IE M 30 $-PARAMETER n = 90 +*IE M 100 $-PARAMETER n = 300 +*IE M 500 $-PARAMETER n = 1500 + IE M 1000 $-PARAMETER n = 3000 +*IE M 3000 $-PARAMETER n = 9000 + IE M 5 $-PARAMETER n = 15 original value + +* N is the number of variables + + IM N M 3 + +* Problem parameters + + RE ALPHA 1.0 + RE BETA 0.125 + RE GAMMA 0.125 + RE DELTA 0.125 + +* K-set 3 + + IE K1 2 + IE K2 1 + IE K3 1 + IE K4 2 + +* Other parameters + + RI RN N + IA N-1 N -1 + I+ 2M M M + + IE 1 1 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + N GA + N GB + N GC + N GD + +CONSTANTS + + DIXMAANO GA -1.0 + +BOUNDS + + FR DIXMAANO 'DEFAULT' + +START POINT + + XV DIXMAANO 'DEFAULT' 2.0 + +ELEMENT TYPE + + EV SQ X + EV SQB X Y + EV SQC X Y + EV 2PR X Y + +ELEMENT USES + +* First group + + DO I 1 N + XT A(I) SQ + ZV A(I) X X(I) + ND + +* Second group + + DO I 1 N-1 + IA I+1 I 1 + XT B(I) SQB + ZV B(I) X X(I) + ZV B(I) Y X(I+1) + ND + +* Third group + + DO I 1 2M + I+ I+M I M + XT C(I) SQC + ZV C(I) X X(I) + ZV C(I) Y X(I+M) + ND + +* Fourth group + + DO I 1 M + I+ I+2M I 2M + XT D(I) 2PR + ZV D(I) X X(I) + ZV D(I) Y X(I+2M) + ND + +GROUP USES + +* First group + + DO I 1 N + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K1 + R* TMP TMP I/N + OD J + R* AI TMP ALPHA + ZE GA A(I) AI + ND + +* Second group + + DO I 1 N-1 + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K2 + R* TMP TMP I/N + OD J + R* BI TMP BETA + ZE GB B(I) BI + ND + +* Third group + + DO I 1 2M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K3 + R* TMP TMP I/N + OD J + R* CI TMP GAMMA + ZE GC C(I) CI + ND + +* Fourth group + + DO I 1 M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K4 + R* TMP TMP I/N + OD J + R* DI TMP DELTA + ZE GD D(I) DI + ND + +OBJECT BOUND + + LO DIXMAANO 0.0 + +* Solution + +*LO SOLTN 1.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS DIXMAANO + +TEMPORARIES + + R F1 + R F2 + R DF2DY + +INDIVIDUALS + +* First type + + T SQ + F X * X + G X X + X + H X X 2.0 + +* Second type + + T SQB + A F1 X * X + A F2 Y + Y * Y + A DF2DY 1.0 + 2.0 * Y + F F1 * F2 * F2 + G X 2.0 * X * F2 * F2 + G Y 2.0 * F1 * F2 * DF2DY + H X X 2.0 * F2 * F2 + H X Y 4.0 * X * DF2DY * F2 + H Y Y 4.0 * F1 * F2 + + H+ 2.0 * F1 * DF2DY * DF2DY + +* Third type + + T SQC + A F1 X * X + A F2 Y**4 + F F1 * F2 + G X 2.0 * X * F2 + G Y 4.0 * F1 * Y**3 + H X X 2.0 * F2 + H X Y 8.0 * X * Y**3 + H Y Y 12.0 * F1 * Y**2 + +* Fourth type + + T 2PR + F X * Y + G X Y + G Y X + H X Y 1.0 + +ENDATA diff --git a/DIXMAANP.SIF b/DIXMAANP.SIF new file mode 100644 index 0000000..9d457d9 --- /dev/null +++ b/DIXMAANP.SIF @@ -0,0 +1,255 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME DIXMAANP + +* Problem : +* ********* +* A variant on the Dixon-Maany test problem (version I) + +* Source: +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* (problem 22) based on + +* L.C.W. Dixon and Z. Maany, +* "A family of test problems with sparse Hessians for unconstrained +* optimization", +* TR 206, Numerical Optimization Centre, Hatfield Polytechnic, 1988. + +* SIF input: Ph. Toint, Dec 1989. +* correction by Ph. Shott, January 1995. +* this version Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* M is equal to the third of the number of variables + +*IE M 5 $-PARAMETER n = 15 original value +*IE M 30 $-PARAMETER n = 90 +*IE M 100 $-PARAMETER n = 300 +*IE M 500 $-PARAMETER n = 1500 + IE M 1000 $-PARAMETER n = 3000 +*IE M 3000 $-PARAMETER n = 9000 + IE M 5 $-PARAMETER n = 15 original value + +* N is the number of variables + + IM N M 3 + +* Problem parameters + + RE ALPHA 1.0 + RE BETA 0.26 + RE GAMMA 0.26 + RE DELTA 0.26 + +* K-set 3 + + IE K1 2 + IE K2 1 + IE K3 1 + IE K4 2 + +* Other parameters + + RI RN N + IA N-1 N -1 + I+ 2M M M + + IE 1 1 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + N GA + N GB + N GC + N GD + +CONSTANTS + + DIXMAANP GA -1.0 + +BOUNDS + + FR DIXMAANP 'DEFAULT' + +START POINT + + XV DIXMAANP 'DEFAULT' 2.0 + +ELEMENT TYPE + + EV SQ X + EV SQB X Y + EV SQC X Y + EV 2PR X Y + +ELEMENT USES + +* First group + + DO I 1 N + XT A(I) SQ + ZV A(I) X X(I) + ND + +* Second group + + DO I 1 N-1 + IA I+1 I 1 + XT B(I) SQB + ZV B(I) X X(I) + ZV B(I) Y X(I+1) + ND + +* Third group + + DO I 1 2M + I+ I+M I M + XT C(I) SQC + ZV C(I) X X(I) + ZV C(I) Y X(I+M) + ND + +* Fourth group + + DO I 1 M + I+ I+2M I 2M + XT D(I) 2PR + ZV D(I) X X(I) + ZV D(I) Y X(I+2M) + ND + +GROUP USES + +* First group + + DO I 1 N + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K1 + R* TMP TMP I/N + OD J + R* AI TMP ALPHA + ZE GA A(I) AI + ND + +* Second group + + DO I 1 N-1 + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K2 + R* TMP TMP I/N + OD J + R* BI TMP BETA + ZE GB B(I) BI + ND + +* Third group + + DO I 1 2M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K3 + R* TMP TMP I/N + OD J + R* CI TMP GAMMA + ZE GC C(I) CI + ND + +* Fourth group + + DO I 1 M + RI RI I + R/ I/N RI RN + RE TMP 1.0 + DO J 1 K4 + R* TMP TMP I/N + OD J + R* DI TMP DELTA + ZE GD D(I) DI + ND + +OBJECT BOUND + + LO DIXMAANP 0.0 + +* Solution + +*LO SOLTN 1.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS DIXMAANP + +TEMPORARIES + + R F1 + R F2 + R DF2DY + +INDIVIDUALS + +* First type + + T SQ + F X * X + G X X + X + H X X 2.0 + +* Second type + + T SQB + A F1 X * X + A F2 Y + Y * Y + A DF2DY 1.0 + 2.0 * Y + F F1 * F2 * F2 + G X 2.0 * X * F2 * F2 + G Y 2.0 * F1 * F2 * DF2DY + H X X 2.0 * F2 * F2 + H X Y 4.0 * X * DF2DY * F2 + H Y Y 4.0 * F1 * F2 + + H+ 2.0 * F1 * DF2DY * DF2DY + +* Third type + + T SQC + A F1 X * X + A F2 Y**4 + F F1 * F2 + G X 2.0 * X * F2 + G Y 4.0 * F1 * Y**3 + H X X 2.0 * F2 + H X Y 8.0 * X * Y**3 + H Y Y 12.0 * F1 * Y**2 + +* Fourth type + + T 2PR + F X * Y + G X Y + G Y X + H X Y 1.0 + +ENDATA diff --git a/ERRINRSM.SIF b/ERRINRSM.SIF new file mode 100644 index 0000000..7f6a582 --- /dev/null +++ b/ERRINRSM.SIF @@ -0,0 +1,138 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME ERRINRSM + +* Problem : +* -------- + +* A variable dimension version of an incorrect version of the +* chained Rosenbrock function (ERRINROS) by Luksan et al. + +* Source: problem 28 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* SIF input: Ph. Toint, Sept 1990. +* this version Nick Gould, June, 2013 + +* classification SUR2-AN-V-0 + +* Number of variables (at most 50) + +*IE N 10 $-PARAMETER +*IE N 25 $-PARAMETER + IE N 50 $-PARAMETER + +* other parameter definitions + + IE 1 1 + IE 2 2 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + DO I 2 N + IA I-1 I -1 + XN SQ(I) X(I-1) 1.0 + XN B(I) X(I) 1.0 + ND + +CONSTANTS + + DO I 2 N + X ERRINRSM B(I) 1.0 + ND + +BOUNDS + + FR ERRINRSM 'DEFAULT' + +START POINT + + DO I 1 N + X ERRINRSM X(I) -1.0 + ND + +ELEMENT TYPE + + EV ETYPE V1 + +ELEMENT USES + + XT 'DEFAULT' ETYPE + + DO I 2 N + ZV ELA(I) V1 X(I) + ND + +GROUP TYPE + + GV L2 GVAR + +GROUP USES + + XT 'DEFAULT' L2 + + DO I 2 N + RI RI I + R( SINI SIN RI + RA ALPHA SINI 1.5 + R* AI2 ALPHA ALPHA + RM AI AI2 16.0 + ZE SQ(I) ELA(I) AI + ND + +OBJECT BOUND + + LO ERRINRSM 0.0 + +* Solution + +*LO SOLTN(10) 6.69463214 +*LO SOLTN(25) 18.4609060 +*LO SOLTN(50) 39.9041540 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS ERRINRSM + +INDIVIDUALS + + T ETYPE + F - V1 ** 2 + G V1 - 2.0 * V1 + H V1 V1 - 2.0 + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS ERRINRSM + +INDIVIDUALS + + T L2 + + F GVAR * GVAR + G GVAR + GVAR + H 2.0 + +ENDATA diff --git a/FLETBV3M.SIF b/FLETBV3M.SIF new file mode 100644 index 0000000..e1ef73f --- /dev/null +++ b/FLETBV3M.SIF @@ -0,0 +1,199 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME FLETBV3M + +* Problem : +* ********* + +* Variant of FLETCBV3, another boundary value problem, by Luksan et al + +* Source: problem 30 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* based on a scaled version of the first problem given by +* R. Fletcher, +* "An optimal positive definite update for sparse Hessian matrices" +* Numerical Analysis report NA/145, University of Dundee, 1992. + +* SIF input: Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* The number of variables is N. + +*IE N 10 $-PARAMETER original value +*IE N 100 $-PARAMETER +*IE N 1000 $-PARAMETER + IE N 5000 $-PARAMETER +*IE N 10000 $-PARAMETER + +* KAPPA a parameter. + + RE KAPPA 1.0 $-PARAMETER +*RE KAPPA 0.0 $-PARAMETER + + RE OBJSCALE 1.0D+8 + +* Define useful parameters + + IE 0 0 + IE 1 1 + IE 2 2 + RE 1.0 1.0 + IA N-1 N -1 + + R/ P 1.0 OBJSCALE + + IA N+1 N 1 + RI RN+1 N+1 + R/ H 1.0 RN+1 + R* H2 H H + R* 1/H2 RN+1 RN+1 + R* KAPPA/H2 1/H2 KAPPA + RM -KAPPA/H2 KAPPA/H2 -1.0 + RM 2/H2 1/H2 2.0 + RA 1+2/H2 2/H2 1.0 + RM -1-2/H2 1+2/H2 -1.0 + R* P*-1-2/H2 1+2/H2 P + + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + N S + +*ZN G(0) 'SCALE' OBJSCALE + XN G(0) X(1) 1.0 + + DO I 1 N-1 + IA I+1 I 1 +*ZN G(I) 'SCALE' OBJSCALE + XN G(I) X(I) 1.0 X(I+1) -1.0 + ND + +*ZN G(N) 'SCALE' OBJSCALE + XN G(N) X(N) 1.0 + + DO I 1 N +*ZN C(I) 'SCALE' OBJSCALE + ZN C(I) + ND + +BOUNDS + + FR FLETBV3M 'DEFAULT' + +START POINT + +*V FLETBV3M 'DEFAULT' 0.0 + + DO I 1 N + RI RI I + R* IH RI H + Z FLETBV3M X(I) IH + ND + +ELEMENT TYPE + + EV SIN V + + EV COS V + EP COS P + +ELEMENT USES + + DO I 1 N + XT C(I) COS + ZV C(I) V X(I) + ZP C(I) P P + ND + + DO I 1 N + XT S(I) SIN + ZV S(I) V X(I) + ND + +GROUP TYPE + + GV HALFL2 GVAR + GP HALFL2 P + +GROUP USES + + DO I 0 N + XT G(I) HALFL2 + ZP G(I) P P + ND + + DO I 1 N + ZE C(I) C(I) -KAPPA/H2 + ND + + DO I 1 N + ZE S S(I) P*-1-2/H2 + ND + + +OBJECT BOUND + +* Solution + +*LO SOLTN ?? + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS FLETBV3M + +TEMPORARIES + + M COS + M SIN + +INDIVIDUALS + + T COS + F P * COS( V ) + G V - P * SIN( V ) + H V V - P * COS( V ) + + T SIN + F 100.0 * SIN( 0.01 * V ) + G V COS( 0.01 * V ) + H V V - 0.01 * SIN( 0.01 * V ) + + + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS FLETBV3M + +INDIVIDUALS + + T HALFL2 + + F 5.0D-1 * P * GVAR * GVAR + G P * GVAR + H P + +ENDATA diff --git a/INDEFM.SIF b/INDEFM.SIF new file mode 100644 index 0000000..60b26c3 --- /dev/null +++ b/INDEFM.SIF @@ -0,0 +1,140 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME INDEFM + +* Problem : +* ********* + +* Variant of INDEF, a nonconvex problem which has an indefinite Hessian +* at the starting point, by Luksan et al + +* Source: problem 37 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* based on the original problem by N. Gould + +* SIF input: Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* The number of variables is N. + + IE N 10 $-PARAMETER +*IE N 50 $-PARAMETER +*IE N 100 $-PARAMETER +*IE N 1000 $-PARAMETER original value + IE N 5000 $-PARAMETER + IE N 10000 $-PARAMETER + IE N 100000 $-PARAMETER + +* The parameter ALPHA controls the indefiniteness. +* Larger values of ALPHA give more indefiniteness. + + RE ALPHA 0.5 $-PARAMETER indefiniteness +*RE ALPHA 1.0 $-PARAMETER +*RE ALPHA 10.0 $-PARAMETER +*RE ALPHA 100.0 $-PARAMETER +*RE ALPHA 1000.0 $-PARAMETER + +* Other parameters + + IE 1 1 + IE 2 2 + IA N-1 N -1 + IA N+1 N 1 + RI RN+1 N+1 + +VARIABLES + + DO I 1 N + X X(I) + ND + +GROUPS + + DO I 1 N + XN SIN(I) X(I) 1.0 + ND + + DO I 2 N-1 + XN COS(I) X(I) 2.0 X(N) -1.0 + XN COS(I) X(1) -1.0 + ND + +BOUNDS + + FR INDEFM 'DEFAULT' + +START POINT + +* start with X(I) = I/N+1. + + DO I 1 N + + RI RI I + R/ T RI RN+1 + ZV INDEF1 X(I) T + + ND + + V INDEF2 'DEFAULT' 1000.0 + +GROUP TYPE + + GV COS GVAR + GP COS ALPHA + GV SIN GVAR + +GROUP USES + + DO I 2 N-1 + XT COS(I) COS + ZP COS(I) ALPHA ALPHA + ND + + DO I 1 N + XT SIN(I) SIN + ND + +OBJECT BOUND + +* Solution + +*LO SOLTN ?? + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS INDEFM + +TEMPORARIES + + M COS + M SIN + +INDIVIDUALS + + T COS + + F ALPHA * COS( GVAR ) + G - ALPHA * SIN( GVAR ) + H - ALPHA * COS( GVAR ) + + T SIN + + F 100.0 * SIN( 0.01 * GVAR ) + G COS( 0.01 * GVAR ) + H - 0.01 * SIN( 0.01 * GVAR ) + +ENDATA + diff --git a/NCB20.SIF b/NCB20.SIF index ca2a24b..4c0acac 100644 --- a/NCB20.SIF +++ b/NCB20.SIF @@ -591,7 +591,7 @@ INDIVIDUALS T 3P F X * Y * Z + 2.0 * Z * Z G X Y * Z - G Y Z * Z + G Y X * Z G Z X * Y + 4.0 * Z H X Y Z H X Z Y diff --git a/SSBRYBND.SIF b/SSBRYBND.SIF new file mode 100644 index 0000000..4e219d6 --- /dev/null +++ b/SSBRYBND.SIF @@ -0,0 +1,298 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME SSBRYBND + +* Problem : +* ********* +* Broyden banded system of nonlinear equations, considered in the +* least square sense. +* NB: scaled version of BRYBND with scaling proposed by Luksan et al. + +* Source: problem 48 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* that is a scaled variant of problem 31 in + +* J.J. More', B.S. Garbow and K.E. Hillstrom, +* "Testing Unconstrained Optimization Software", +* ACM Transactions on Mathematical Software, vol. 7(1), pp. 17-41, 1981. + +* See also Buckley#73 (p. 41) and Toint#18 + +* SIF input: Ph. Toint and Nick Gould, Nov 1997. +* this version Nick Gould, June, 2013 + +* classification SUR2-AN-V-0 + +* N is the number of equations and variables (variable). + +*IE N 10 $-PARAMETER +*IE N 50 $-PARAMETER +*IE N 100 $-PARAMETER +*IE N 500 $-PARAMETER +*IE N 1000 $-PARAMETER original value + IE N 5000 $-PARAMETER + +* Define some problem's parameters (see Buckley) +* Restriction: LB + 1 + UB .le. N + + RE ONE 1.0 + RE KAPPA1 2.0 + RE KAPPA2 5.0 + RE KAPPA3 1.0 + IE LB 5 + IE UB 1 + RI RN N + RA RN-1 RN -1 + +* ratio of smallest to largest scale factors will be exp(scal) + + RE SCAL 6.0 + +* Define useful parameters + + IE 1 1 + IM MLB LB -1 + IM MUB UB -1 + IA LB+1 LB 1 + I+ N-UB N MUB + IA N-UB-1 N-UB -1 + RM -KAPPA3 KAPPA3 -1.0 + +VARIABLES + + DO I 1 N + IA I-1 I -1 + RI RI-1 I-1 + R/ RAT RI-1 RN-1 + R* ARG RAT SCAL + A( SCALE(I) EXP ARG + X X(I) + ND + +GROUPS + +* Upper left corner + + DO I 1 LB + + IA I-1 I -1 + IA I+1 I 1 + I+ I+UB I UB + + DO J 1 I-1 + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + OD J + + A* KAP KAPPA1 SCALE(I) + ZN G(I) X(I) KAP + + DO J I+1 I+UB + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + + ND + +* Main (middle) part + + DO I LB+1 N-UB-1 + + I+ I-LB I MLB + IA I-1 I -1 + IA I+1 I 1 + I+ I+UB I UB + + DO J I-LB I-1 + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + OD J + + A* KAP KAPPA1 SCALE(I) + ZN G(I) X(I) KAP + + DO J I+1 I+UB + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + + ND + +* Lower right corner + + DO I N-UB N + + I+ I-LB I MLB + IA I-1 I -1 + IA I+1 I 1 + + DO J I-LB I-1 + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + OD J + + A* KAP KAPPA1 SCALE(I) + ZN G(I) X(I) KAP + + DO J I+1 N + A* KAP -KAPPA3 SCALE(J) + ZN G(I) X(J) KAP + + ND + +BOUNDS + + FR SSBRYBND 'DEFAULT' + +START POINT + + DO I 1 N + A/ DIV ONE SCALE(I) + ZV SSBRYBND X(I) DIV + ND + +ELEMENT TYPE + + EV SQ V + EP SQ P + EV CB V + EP CB P + +ELEMENT USES + + XT 'DEFAULT' CB + + DO I 1 N + XT E(I) SQ + ZV E(I) V X(I) + ZP E(I) P SCALE(I) + XT Q(I) CB + ZV Q(I) V X(I) + ZP Q(I) P SCALE(I) + ND + +GROUP TYPE + + GV L2 GVAR + +GROUP USES + + XT 'DEFAULT' L2 + +* Upper left corner + + DO I 1 LB + + IA I-1 I -1 + IA I+1 I 1 + I+ I+UB I UB + + DO J 1 I-1 + ZE G(I) E(J) -KAPPA3 + OD J + + ZE G(I) Q(I) KAPPA2 + + DO J I+1 I+UB + ZE G(I) E(J) -KAPPA3 + + ND + +* Main (middle) part + + DO I LB+1 N-UB-1 + + I+ I-LB I MLB + IA I-1 I -1 + IA I+1 I 1 + I+ I+UB I UB + + DO J I-LB I-1 + ZE G(I) Q(J) -KAPPA3 + OD J + + ZE G(I) E(I) KAPPA2 + + DO J I+1 I+UB + ZE G(I) E(J) -KAPPA3 + + ND + +* Lower right corner + + DO I N-UB N + + I+ I-LB I MLB + IA I-1 I -1 + IA I+1 I 1 + + DO J I-LB I-1 + ZE G(I) E(J) -KAPPA3 + OD J + + ZE G(I) Q(I) KAPPA2 + + DO J I+1 N + ZE G(I) E(J) -KAPPA3 + + ND + +OBJECT BOUND + +* Solution + +*LO SOLTN 0.0 + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS SSBRYBND + +TEMPORARIES + + R PP + +INDIVIDUALS + + T SQ + A PP P * P + F PP * V * V + G V PP * ( V + V ) + H V V 2.0 * PP + + T CB + A PP P * P * P + F PP * V * V * V + G V 3.0 * PP * V * V + H V V 6.0 * PP * V + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS SSBRYBND + +INDIVIDUALS + + T L2 + + F GVAR * GVAR + G GVAR + GVAR + H 2.0 + +ENDATA + + + diff --git a/SSCOSINE.SIF b/SSCOSINE.SIF new file mode 100644 index 0000000..d8c8aba --- /dev/null +++ b/SSCOSINE.SIF @@ -0,0 +1,165 @@ +*************************** +* SET UP THE INITIAL DATA * +*************************** + +NAME SSCOSINE + +* Problem : +* ********* + +* Another function with nontrivial groups and +* repetitious elements. +* NB: scaled version of COSINE by Luksan et al. + +* Source: problem 50 in +* L. Luksan, C. Matonoha and J. Vlcek +* Modified CUTE problems for sparse unconstraoined optimization +* Technical Report 1081 +* Institute of Computer Science +* Academy of Science of the Czech Republic + +* that is a scaled variant of +* N. Gould, private communication. + +* SIF input: N. Gould, Nov 1997 +* this version Nick Gould, June, 2013 + +* classification OUR2-AN-V-0 + +* number of variables + +*IE N 10 $-PARAMETER +*IE N 100 $-PARAMETER +*IE N 1000 $-PARAMETER original value + IE N 5000 $-PARAMETER +*IE N 10000 $-PARAMETER + IE N 100000 $-PARAMETER +*IE N 1000000 $-PARAMETER + IE N 10 $-PARAMETER + +* ratio of smallest to largest scale factors will be exp(scal) + + RE SCAL 6.0 + +* other parameter definitions + + IE 1 1 + IA N-1 N -1 + RI RN-1 N-1 + RM -RN-1 RN-1 -1.0 + + RE ONE 1.0 + RI RN N + RA RN-1 RN -1 + +VARIABLES + + DO I 1 N + IA I-1 I -1 + RI RI-1 I-1 + R/ RAT RI-1 RN-1 + R* ARG RAT SCAL + A( S(I) EXP ARG + X X(I) + ND + +GROUPS + +* objective function + + DO I 1 N-1 + IA I+1 I 1 + AM MULT S(I+1) -0.5 + ZN G(I) X(I+1) MULT + OD + +BOUNDS + + FR SSCOSINE 'DEFAULT' + +START POINT + + DO I 1 N + A/ DIV ONE S(I) + ZV SSCOSINE X(I) DIV + ND + +ELEMENT TYPE + + EV SQ V + EP SQ P + +ELEMENT USES + + DO I 1 N-1 + + XT E(I) SQ + ZV E(I) V X(I) + ZP E(I) P S(I) + + ND + +GROUP TYPE + + GV COS GVAR + +GROUP USES + + DO I 1 N-1 + XT G(I) COS + XE G(I) E(I) + ND + +OBJECT BOUND + + XL SSCOSINE -RN-1 + +* Solution + +*LO SOLTN ??? + +ENDATA + +*********************** +* SET UP THE FUNCTION * +* AND RANGE ROUTINES * +*********************** + +ELEMENTS SSCOSINE + +TEMPORARIES + + R PP + +INDIVIDUALS + + T SQ + A PP P * P + F PP * V * V + G V PP * ( V + V ) + H V V 2.0 * PP + +ENDATA + +********************* +* SET UP THE GROUPS * +* ROUTINE * +********************* + +GROUPS SSCOSINE + +TEMPORARIES + + R COSG + M SIN + M COS + +INDIVIDUALS + + T COS + A COSG COS( GVAR ) + F COSG + G - SIN( GVAR ) + H - COSG + +ENDATA diff --git a/mastsif.html b/mastsif.html index 79735f8..0a7c612 100644 --- a/mastsif.html +++ b/mastsif.html @@ -1230,6 +1230,12 @@

Individual test problems

gzipped file NOR2-AN- V- V + CHNRSNBM + ascii file, + +gzipped file + SUR2-AN- V- 0 + CKOEHELB ascii file, @@ -1692,6 +1698,30 @@

Individual test problems

gzipped file OUR2-AN-V-0 + DIXMAANM + ascii file, + +gzipped file + OUR2-AN- V- 0 + + DIXMAANN + ascii file, + +gzipped file + OUR2-AN- V- 0 + + DIXMAANO + ascii file, + +gzipped file + OUR2-AN- V- 0 + + DIXMAANP + ascii file, + +gzipped file + OUR2-AN- V- 0 + DIXON3DQ ascii file, @@ -2058,6 +2088,12 @@

Individual test problems

gzipped file SUR2-AN-V-0 + ERRINRSM + ascii file, + +gzipped file + SUR2-AN- V- 0 + EXPFIT ascii file, @@ -2142,6 +2178,12 @@

Individual test problems

gzipped file QLR2-RN-34501-58825 + FLETBV3M + ascii file, + +gzipped file + OUR2-AN- V- 0 + FLETCBV2 ascii file, @@ -2154,6 +2196,7 @@

Individual test problems

gzipped file OUR2-AN-V-0 + FLETCHBV ascii file, @@ -3666,6 +3709,12 @@

Individual test problems

gzipped file OUR2-AN-V-0 + INDEFM + ascii file, + +gzipped file + OUR2-AN- V- 0 + INTEGREQ ascii file, @@ -6180,12 +6229,24 @@

Individual test problems

gzipped file SUR2-AN-V-0 + SSBRYBND + ascii file, + +gzipped file + SUR2-AN- V- 0 + SSC ascii file, gzipped file OXR2-MY-V-0 + SSCOSINE + ascii file, + +gzipped file + OUR2-AN- V- 0 + SSEBLIN ascii file, @@ -6926,3 +6987,4 @@

Individual test problems

+