wassname/simpeg
1
0
Fork 0
mirror of https://github.com/wassname/simpeg.git synced 2026-08-03 13:10:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

Implement iterative solver BiCGStab + Jacobi Preconditionner (success!!)

Browse Source
This commit is contained in:
D Fournier
2015-12-11 11:32:03 -08:00
parent 3d52778274
commit 2aa523f0ad
13 changed files with 619 additions and 939 deletions
Download Patch File Download Diff File Expand all files Collapse all files
simpegDCIP/Dev/Inv2D/dcinv2d.log
+203 -181
View File
@@ -14,7 +14,7 @@ Distributed by:
Mira Geoscience Ltd.
DCINV2D started on:12/10/2015 14:20:13
DCINV2D started on:12/11/2015 11:26:38
Reading input file: dcinv2d.inp
@@ -35,17 +35,17 @@ USE_MREF TRUE
maximum # of iterations: 100
data were read from: FWR_3D_2_2D.dat
# of current locations: 10
# of data: 54
# of current locations: 9
# of data: 45
chifact: 1.00000E+00
target misfit: 5.40000E+01
target misfit: 4.50000E+01
mesh was read from: Mesh_2D.msh
# of cells: 50 x 30
total # of cells: 1500
# of active cells: 1500
# of cells: 47 x 30
total # of cells: 1410
# of active cells: 1410
# of unique data locations: 10
# of unique data locations: 9
# of wave values: 13
2.5000E-04 4.9901E-04 9.9606E-04 1.9882E-03 3.9685E-03 7.9213E-03 1.5811E-02 3.1560E-02 6.2996E-02 1.2574E-01 2.5099E-01 5.0099E-01 1.0000E+00
@@ -57,12 +57,12 @@ corresponding alpha (a_s, a_x, a_z): ( 1.56250E-04, 1.0000E+00, 1.0000E+00)
Using basis vectors and SVD.
reference model will be used in the derivative terms.
number of basis vectors: 19 + 3 + 1 = 23
number of basis vectors: 17 + 3 + 1 = 21
init cpu time: 0:00:00.04
init cpu time: 0:00:00.06
initial misfit = 2.87480E+04
initial misfit = 5.86595E+04
init. model norm = 0.00000E+00
norm comp Ws = 0.00000E+00
norm comp Wx = 0.00000E+00
@@ -73,263 +73,263 @@ Iteration 1
beta vs. misfit:
beta misfit
1.05827E+04 7.91147E+03
2.11653E+04 9.97641E+03
5.29133E+04 1.42234E+04
5.43732E+04 1.43748E+04
1.14011E+04 1.55434E+04
2.28022E+04 1.97701E+04
5.70054E+04 2.80956E+04
6.37666E+04 2.93408E+04
chosen beta = 5.43657E+04
target misfit = 1.43740E+04
achieved misfit = 1.43740E+04
model norm = 1.00462E-01
misfit change = 5.00000E-01
chosen beta = 6.37044E+04
target misfit = 2.93298E+04
achieved misfit = 2.93298E+04
model norm = 1.78131E-01
misfit change = 4.99999E-01
model norm change = 0.00000E+00
norm comp Ws = 5.19270E-02
norm comp Wx = 2.47955E-02
norm comp Wz = 2.37393E-02
norm comp Ws = 9.36044E-02
norm comp Wx = 5.21598E-02
norm comp Wz = 3.23664E-02
iter cpu time: 0:00:00.64
iter cpu time: 0:00:00.54
Iteration 2
beta vs. misfit:
beta misfit
8.89064E+03 7.03214E+03
1.35914E+04 8.23443E+03
2.71829E+04 1.09665E+04
1.13370E+04 1.41481E+04
1.59261E+04 1.63158E+04
3.18522E+04 2.20755E+04
chosen beta = 9.42699E+03
target misfit = 7.18700E+03
achieved misfit = 7.18162E+03
model norm = 4.48909E-01
misfit change = 5.00374E-01
model norm change = 3.46846E+00
chosen beta = 1.23494E+04
target misfit = 1.46649E+04
achieved misfit = 1.46622E+04
model norm = 7.62402E-01
misfit change = 5.00094E-01
model norm change = 3.28002E+00
norm comp Ws = 1.84955E-01
norm comp Wx = 1.65870E-01
norm comp Wz = 9.80840E-02
norm comp Ws = 3.10519E-01
norm comp Wx = 3.26120E-01
norm comp Wz = 1.25764E-01
iter cpu time: 0:00:00.40
iter cpu time: 0:00:00.17
Iteration 3
beta vs. misfit:
beta misfit
1.30271E+03 3.38723E+03
2.35675E+03 4.32491E+03
4.71349E+03 5.60434E+03
3.08734E+03 7.25168E+03
6.17468E+03 1.04745E+04
chosen beta = 1.50085E+03
target misfit = 3.59081E+03
achieved misfit = 3.60067E+03
model norm = 1.47576E+00
misfit change = 4.98627E-01
model norm change = 2.28744E+00
chosen beta = 3.15138E+03
target misfit = 7.33109E+03
achieved misfit = 7.33827E+03
model norm = 1.97756E+00
misfit change = 4.99510E-01
model norm change = 1.59385E+00
norm comp Ws = 4.08324E-01
norm comp Wx = 8.20762E-01
norm comp Wz = 2.46674E-01
norm comp Ws = 5.74759E-01
norm comp Wx = 1.07019E+00
norm comp Wz = 3.32607E-01
iter cpu time: 0:00:00.49
iter cpu time: 0:00:00.14
Iteration 4
beta vs. misfit:
beta misfit
3.75212E+02 1.34820E+03
7.50424E+02 2.29809E+03
7.87844E+02 2.34933E+03
1.57569E+03 4.17369E+03
chosen beta = 5.46411E+02
target misfit = 1.80034E+03
achieved misfit = 1.83379E+03
model norm = 3.31412E+00
misfit change = 4.90710E-01
model norm change = 1.24570E+00
chosen beta = 1.34889E+03
target misfit = 3.66914E+03
achieved misfit = 3.69116E+03
model norm = 3.63935E+00
misfit change = 4.96999E-01
model norm change = 8.40325E-01
norm comp Ws = 6.83124E-01
norm comp Wx = 2.07548E+00
norm comp Wz = 5.55519E-01
norm comp Ws = 9.31633E-01
norm comp Wx = 2.00929E+00
norm comp Wz = 6.98426E-01
iter cpu time: 0:00:00.20
iter cpu time: 0:00:00.43
Iteration 5
beta vs. misfit:
beta misfit
1.36603E+02 5.05265E+02
2.73206E+02 9.48634E+02
3.37223E+02 1.15980E+03
6.74446E+02 1.95927E+03
chosen beta = 2.63165E+02
target misfit = 9.16894E+02
achieved misfit = 9.16736E+02
model norm = 5.54836E+00
misfit change = 5.00086E-01
model norm change = 6.74157E-01
chosen beta = 6.23201E+02
target misfit = 1.84558E+03
achieved misfit = 1.83831E+03
model norm = 5.55702E+00
misfit change = 5.01969E-01
model norm change = 5.26927E-01
norm comp Ws = 8.54281E-01
norm comp Wx = 3.70325E+00
norm comp Wz = 9.90823E-01
norm comp Ws = 1.39116E+00
norm comp Wx = 2.97085E+00
norm comp Wz = 1.19501E+00
iter cpu time: 0:00:00.17
iter cpu time: 0:00:00.13
Iteration 6
beta vs. misfit:
beta misfit
6.57912E+01 2.23512E+02
1.31582E+02 4.05775E+02
1.51607E+02 4.68975E+02
1.55800E+02 7.04026E+02
3.11600E+02 1.03301E+03
chosen beta = 1.48251E+02
target misfit = 4.58368E+02
achieved misfit = 4.58273E+02
model norm = 7.68792E+00
misfit change = 5.00103E-01
model norm change = 3.85621E-01
chosen beta = 2.52300E+02
target misfit = 9.19156E+02
achieved misfit = 9.06295E+02
model norm = 7.76502E+00
misfit change = 5.06996E-01
model norm change = 3.97336E-01
norm comp Ws = 1.03639E+00
norm comp Wx = 5.25876E+00
norm comp Wz = 1.39277E+00
norm comp Ws = 1.93177E+00
norm comp Wx = 4.15601E+00
norm comp Wz = 1.67723E+00
iter cpu time: 0:00:00.29
iter cpu time: 0:00:00.14
Iteration 7
beta vs. misfit:
beta misfit
3.70626E+01 1.66504E+02
7.41253E+01 2.51726E+02
5.45665E+01 4.43592E+02
6.30749E+01 4.61153E+02
1.26150E+02 5.89041E+02
chosen beta = 6.33124E+01
target misfit = 2.29137E+02
achieved misfit = 2.25244E+02
model norm = 1.00149E+01
misfit change = 5.08495E-01
model norm change = 3.02681E-01
chosen beta = 5.90843E+01
target misfit = 4.53148E+02
achieved misfit = 4.52929E+02
model norm = 1.10434E+01
misfit change = 5.00242E-01
model norm change = 4.22194E-01
norm comp Ws = 1.04755E+00
norm comp Wx = 6.91957E+00
norm comp Wz = 2.04778E+00
norm comp Ws = 2.44126E+00
norm comp Wx = 6.33488E+00
norm comp Wz = 2.26722E+00
iter cpu time: 0:00:00.34
iter cpu time: 0:00:00.15
Iteration 8
beta vs. misfit:
beta misfit
1.58281E+01 1.02052E+02
3.16562E+01 1.24062E+02
1.76205E+00 2.33067E+02
3.87651E+00 2.33770E+02
8.52832E+00 2.36313E+02
1.47711E+01 2.41686E+02
2.95422E+01 2.66924E+02
chosen beta = 2.24560E+01
target misfit = 1.12622E+02
achieved misfit = 1.09408E+02
model norm = 1.25018E+01
misfit change = 5.14267E-01
model norm change = 2.48315E-01
chosen beta = 3.87651E+00
target misfit = 2.26464E+02
achieved misfit = 2.33770E+02
model norm = 1.57679E+01
misfit change = 4.83870E-01
model norm change = 4.27818E-01
norm comp Ws = 1.21846E+00
norm comp Wx = 9.04398E+00
norm comp Wz = 2.23932E+00
norm comp Ws = 3.91956E+00
norm comp Wx = 8.05061E+00
norm comp Wz = 3.79775E+00
iter cpu time: 0:00:00.17
iter cpu time: 0:00:00.54
Iteration 9
beta vs. misfit:
beta misfit
1.15992E+00 7.28154E+01
2.55182E+00 7.30977E+01
5.61400E+00 7.44520E+01
1.12280E+01 7.91248E+01
9.69127E-01 1.25628E+02
1.93825E+00 1.25177E+02
3.87651E+00 1.26747E+02
chosen beta = 2.55182E+00
target misfit = 5.47041E+01
achieved misfit = 7.30977E+01
model norm = 1.46799E+01
misfit change = 3.31881E-01
model norm change = 1.74228E-01
chosen beta = 1.93825E+00
target misfit = 1.16885E+02
achieved misfit = 1.25177E+02
model norm = 2.02946E+01
misfit change = 4.64529E-01
model norm change = 2.87081E-01
norm comp Ws = 1.27978E+00
norm comp Wx = 1.03955E+01
norm comp Wz = 3.00464E+00
norm comp Ws = 5.14918E+00
norm comp Wx = 1.04854E+01
norm comp Wz = 4.66004E+00
iter cpu time: 0:00:00.17
iter cpu time: 0:00:00.14
Iteration 10
beta vs. misfit:
beta misfit
1.39261E+00 4.80711E+01
1.88512E+00 4.83598E+01
4.71280E+00 5.04589E+01
1.17820E+01 5.88251E+01
4.84563E-01 8.59120E+01
9.69127E-01 8.56434E+01
1.93825E+00 8.63895E+01
chosen beta = 7.06669E+00
target misfit = 5.40000E+01
achieved misfit = 5.27458E+01
model norm = 1.48886E+01
misfit change = 2.78420E-01
model norm change = 1.42172E-02
chosen beta = 9.69127E-01
target misfit = 6.25885E+01
achieved misfit = 8.56434E+01
model norm = 2.27540E+01
misfit change = 3.15821E-01
model norm change = 1.21187E-01
norm comp Ws = 1.42273E+00
norm comp Wx = 1.05086E+01
norm comp Wz = 2.95727E+00
norm comp Ws = 6.37929E+00
norm comp Wx = 1.05750E+01
norm comp Wz = 5.79974E+00
iter cpu time: 0:00:00.39
iter cpu time: 0:00:00.13
Iteration 11
beta vs. misfit:
beta misfit
7.23472E+00 4.12130E+01
7.40675E+00 4.14256E+01
1.85169E+01 5.95740E+01
2.67558E-01 5.36200E+01
5.09212E-01 5.36850E+01
chosen beta = 1.44535E+01
target misfit = 5.40000E+01
achieved misfit = 5.20734E+01
model norm = 1.45035E+01
misfit change = 1.27476E-02
model norm change = -2.58673E-02
chosen beta = 5.09212E-01
target misfit = 4.50000E+01
achieved misfit = 5.36850E+01
model norm = 2.65060E+01
misfit change = 3.73156E-01
model norm change = 1.64894E-01
norm comp Ws = 1.38232E+00
norm comp Wx = 1.04192E+01
norm comp Wz = 2.70198E+00
norm comp Ws = 7.06440E+00
norm comp Wx = 1.17778E+01
norm comp Wz = 7.66382E+00
iter cpu time: 0:00:00.26
iter cpu time: 0:00:00.12
Iteration 12
beta vs. misfit:
beta misfit
1.49882E+01 5.14497E+01
1.55427E+01 5.22662E+01
1.67582E+01 5.41032E+01
3.57781E-01 4.27021E+01
4.26833E-01 4.27304E+01
1.06708E+00 4.32906E+01
2.66771E+00 4.58497E+01
chosen beta = 1.66886E+01
target misfit = 5.40000E+01
achieved misfit = 5.39963E+01
model norm = 1.42758E+01
misfit change = -3.69261E-02
model norm change = -1.56990E-02
chosen beta = 1.97938E+00
target misfit = 4.50000E+01
achieved misfit = 4.46230E+01
model norm = 2.75984E+01
misfit change = 1.68800E-01
model norm change = 4.12137E-02
norm comp Ws = 1.40497E+00
norm comp Wx = 1.00935E+01
norm comp Wz = 2.77734E+00
norm comp Ws = 7.60316E+00
norm comp Wx = 1.17705E+01
norm comp Wz = 8.22479E+00
iter cpu time: 0:00:00.17
iter cpu time: 0:00:00.51
Target misfit achieved. Minimizing model norm.
@@ -339,30 +339,52 @@ Iteration 13
beta vs. misfit:
beta misfit
1.66898E+01 5.35819E+01
1.66909E+01 5.35835E+01
1.69863E+01 5.40047E+01
1.99610E+00 3.93953E+01
2.01297E+00 3.94209E+01
5.03242E+00 4.59121E+01
chosen beta = 1.69830E+01
target misfit = 5.40000E+01
achieved misfit = 5.39999E+01
model norm = 1.42353E+01
misfit change = -6.75492E-05
model norm change = -2.83385E-03
chosen beta = 4.46061E+00
target misfit = 4.50000E+01
achieved misfit = 4.44143E+01
model norm = 2.65756E+01
misfit change = 4.67734E-03
model norm change = -3.70618E-02
norm comp Ws = 1.36827E+00
norm comp Wx = 1.00376E+01
norm comp Wz = 2.82945E+00
norm comp Ws = 7.28818E+00
norm comp Wx = 1.16493E+01
norm comp Wz = 7.63807E+00
iter cpu time: 0:00:00.18
iter cpu time: 0:00:00.15
Iteration 14
beta vs. misfit:
beta misfit
4.51943E+00 4.42228E+01
4.57903E+00 4.43808E+01
4.81808E+00 4.50262E+01
chosen beta = 4.80820E+00
target misfit = 4.50000E+01
achieved misfit = 4.49992E+01
model norm = 2.59337E+01
misfit change = -1.31690E-02
model norm change = -2.41533E-02
norm comp Ws = 7.05571E+00
norm comp Wx = 1.14342E+01
norm comp Wz = 7.44373E+00
iter cpu time: 0:00:00.15
Exit at convergence.
Iterations performed: 13
Iterations performed: 14
total cpu time: 0:00:03.96
total cpu time: 0:00:03.57
DCINV2D ended on:12/10/2015 14:20:17
DCINV2D ended on:12/11/2015 11:26:42