From b8f8dcbb150e803f88725844f73afbce0d91a3c0 Mon Sep 17 00:00:00 2001
From: Max Schweikart <hello@maxschweik.art>
Date: Fri, 10 Feb 2023 14:03:26 +0100
Subject: [PATCH] chore: remove compiled GAMS files

---
 .gitignore             |   1 +
 gams/maxCut/maxcut.lst | 531 -----------------------------------------
 gams/maxcut/maxcut.lst | 427 ---------------------------------
 gams/sat/sat.lst       | 221 -----------------
 gradlew                |   0
 5 files changed, 1 insertion(+), 1179 deletions(-)
 delete mode 100644 gams/maxCut/maxcut.lst
 delete mode 100644 gams/maxcut/maxcut.lst
 delete mode 100644 gams/sat/sat.lst
 mode change 100644 => 100755 gradlew

diff --git a/.gitignore b/.gitignore
index 3f19aabc..45650d4d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -40,3 +40,4 @@ out/
 ### Toolbox ###
 /jobs/
 gamslice.txt
+*.lst
diff --git a/gams/maxCut/maxcut.lst b/gams/maxCut/maxcut.lst
deleted file mode 100644
index 32227bb8..00000000
--- a/gams/maxCut/maxcut.lst
+++ /dev/null
@@ -1,531 +0,0 @@
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 1
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-C o m p i l a t i o n
-
-
-   2   
-      Let G(N, E) denote a graph. A cut is a partition of the vertices N
-      into two sets S and T. Any edge (u,v) in E with u in S and v in T is
-      said to be crossing the cut and is a cut edge. The size of the cut is
-      defined to be sum of weights of the edges crossing the cut.
-       
-      This model presents a simple MIP formulation of the problem that is
-      seeded with a solution from the Goemans/Williamson randomized
-      approximation algorithm based on a semidefinite programming
-      relaxation. By default CSDP is used to solve the SDP.
-      Use --SDPSOLVER=MOSEK to switch to Mosek.
-       
-      The MaxCut instance tg20_7777 is available from the Biq Mac Library
-      and comes from applications in statistical physics.
-       
-       
-      Wiegele A., Biq Mac Library - Binary Quadratic and Max Cut Library.
-      http://biqmac.uni-klu.ac.at/biqmaclib.html
-       
-      Goemans M.X., and Williamson, D.P., Improved Approximation Algorithms
-      for Maximum Cut and Satisfiability Problems Using Semidefinite
-      Programming. Journal of the ACM 42 (1995), 1115-1145.
-      http://www-math.mit.edu/~goemans/PAPERS/maxcut-jacm.pdf
-       
-      Keywords: mixed integer linear programming, approximation algorithms,
-                convex optimization, randomized algorithms, maximum cut problem,
-                mathematics
-  31   
-  39   
-  40  Set n 'nodes';
-  41   
-  42  Alias (n,i,j);
-  43   
-  44  Parameter w(i,j) 'edge weights';
-  45   
-  46  Set e(i,j) 'edges';
-  47   
-  48  Set S(n), T(n), bestS(n);
-  49   
-  50  Scalar
-  51      wS      weight of cut S / -inf /
-  52      maxwS   best weight / -inf /
-  53      mingapS min gap / inf /;
-  54   
-  55  Scalar SDPRelaxation / inf /;
-  56   
-  57  parameter rep(*) / 'value of cut'              -inf
-  58                     'value of best known bound' +inf
-  59                     'relative gap'              NA   /;
-  60   
-  61   
-  68   
-  69  * We want all edges to be i-j with i<j;
-  70  e(i,j)    = ord(i) < ord(j);
-  71  w(e(i,j)) = w(i,j) + w(j,i);
-  72  w(i,j)$(not e(i,j)) = 0;
-  73   
-  74  option e < w;
-  75  *option e:0:0:1, w:8:0:1; display n,e,w;
-  76   
-  77  * Simple MIP model
-  78  Variable
-  79     x(n)     'decides on what side of the cut'
-  80     cut(i,j) 'edge is in the cut'
-  81     z        'objective';
-  82   
-  83  Binary Variable x;
-  84   
-  85  Equation obj, xor1(i,j), xor2(i,j), xor3(i,j), xor4(i,j);
-  86   
-  87  obj..          z      =e= sum(e, w(e)*cut(e));
-  88   
-  89  xor1(e(i,j)).. cut(e) =l= x(i) + x(j);
-  90   
-  91  xor2(e(i,j)).. cut(e) =l= 2 - x(i) - x(j);
-  92   
-  93  xor3(e(i,j)).. cut(e) =g= x(i) - x(j);
-  94   
-  95  xor4(e(i,j)).. cut(e) =g= x(j) - x(i);
-  96   
-  97  Model maxcut / all /;
-  98   
-  99   
- 100  * QUBO Model
- 101  equation defQUBO;
- 102  defQUBO.. z =e= sum(e(i,j), w(e)*(sqr(x(i)) + sqr(x(j)) -2*x(i)*x(j)));
- 103  model qubo / defQUBO /;
- 104   
- 105   
-      Set up the SDP
-         max W*Y s.t. Y_ii = 1, Y positive semidefinite (psd)
-      We need to pass on the dual to csdp
-         min x1 + x2 + ... + xn s.t. X = F1*x1 + F2*x2 + ... + Fn*xn - W, X psd
-      with F_i = 1 for F_ii and 0 otherwise
- 114   
- 115   
- 116  Parameter L(i,j) 'Cholesky factor of Y';
- 117   
- 119   
- 120  Variable Y(i,j)    'PSDMATRIX';
- 121  Variable sdpobj    'objective function variable';
- 122  Equation sdpobjdef 'objective function W*Y';
- 123  sdpobjdef.. sum(e(i,j),w(i,j)*(Y(i,j)+Y(j,i))/2.0) + sum((i,j),eps*Y(i,j)) =E= sdpobj;
- 124  Y.fx(i,i) = 1.0;
- 125  Model sdp / sdpobjdef /;
- 126   
- 127  option lp = MOSEK;
- 128  sdp.limrow   = 0;
- 129  sdp.limcol   = 0;
- 130  sdp.solprint = 0;
- 131  sdp.reslim   = 60-timeelapsed;
- 132  Solve sdp min sdpobj using lp;
- 133   
- 134  Parameter Yl(i,j)   'level values of Y as parameter';
- 135  Yl(i,j) = Y.l(i,j);
-LIBINCLUDE C:\Program Files (x86)\GAMS\42\inclib\linalg.gms
- 137  *
- 138  * LibInclude file to provide functionality of old matrix utilities cholesky via EmbeddedCode and numpy.linalg
- 139  * -c runs this at compile time, -e at execution time (default)
- 140  *
- 141  * Usage: $libInclude linalg [-c|-e] cholesky i A L
- 142  *        This calculates the Cholesky decomposition of a symmetric positive definite matrix A: A = LL^t
- 143  *        The matrix A is indexed over A(i,i)
- 144  *
- 145  * Usage: $libInclude linalg [-c|-e] eigenvalue i A AVal
- 146  *        This calculates the Eigenvalues of a symmetric positive definite matrix.
- 147  *        The matrix A is indexed over A(i,i). AVal(i) is indexed over i.
- 148  *
- 149  * Usage: $libInclude linalg [-c|-e] eigenvector i A AVal AVec
- 150  *        This calculates the Eigenvalues and Eigenvecors of a symmetric positive definite matrix.
- 151  *        The matrices A and AVec are indexed over (i,i). AVal(i) is indexed over i.
- 152  *
- 153  * Usage: $libInclude linalg [-c|-e] invert i A AInv
- 154  *        This calculates the inverse of a square matrix A: A*AInv = I
- 155  *        The matrices A and AInv are indexed over (i,i)
- 156  *
- 157  * Usage: $libInclude linalg [-c|-e] ols [-info=infoSym] [-cfival=val] [-intercept=0,1,2] [-CDFStudentT=fname] [-iCDFStudentT=fname] [-rcond=val] i p A y estimate
- 158  *        This estimates  the unknown parameters in a linear regression model.
- 159  *        The set i are the observations, the set p are the estimates. The matrices
- 160  *        A(i,p) contains the explanatory variable and y(i) the dependent variable.
- 161  *        On return the symbol estimate(p) will contain the estimated statistical
- 162  *        coefficients. The following parameters are available:
- 163  *          -intercept=0,1,2      For 0 no constant term or intercept will be added to
- 164  *                                the problem. For 1 a constant term will always be added.
- 165  *                                For 2 (default) the algorithm will add a constant term
- 166  *                                only if there is no data column with all ones in the matrix A.
- 167  *          -cfival=val           The confidence interval value. Default is 0.95. This impacts
- 168  *                                the calculation of confint_lo and confint_up.
- 169  *          -CDFStudentT=fname    The name of the cumulative distribution function of the
- 170  *                                StudentT distribution. The GAMS extrinisc function library
- 171  *                                stodclib provides this function. The default is CDFStudentT.
- 172  *                                The CDFStudentT is required to calculate pval.
- 173  *          -iCDFStudentT=fname   The name of the inverse cumulative distribution function of the
- 174  *                                StudentT distribution. The GAMS extrinisc function library
- 175  *                                stodclib provides this function. The default is iCDFStudentT.
- 176  *                                The iCDFStudentT is required to calculate confint_lo/up.
- 177  *          -rcond=val            Cut-off ratio for small singular values of A used as argument
- 178  *                                rcond to np.linalg.lstsq. Default is -1.
- 179  *          -mergetype=val        Determines if the symbols will be merged or replaced when loading
- 180  *                                into GAMS. Possible values are DEFAULT, REPLACE, and MERGE.
- 181  *                                Capitalization of this argument is important.
- 182  *                                See https://www.gams.com/latest/docs/UG_EmbeddedCode.html#UG_EmbeddedCode_Python
- 183  *        Additonal regression statistics can be requested via -info=infoSym. -info without the
- 184  *        infoSym, the name of info is used as a symbol name, e.g. -df sets GAMS scalar df.
- 185  *        The following statistics are available:
- 186  *          info        domain description
- 187  *          confint_lo  p      confidence interval (lower bound)
- 188  *          confint_up  p      confidence interval (upper bound)
- 189  *          covar       p,p    variance-covariance matrix
- 190  *          df                 degrees of freedom
- 191  *          fitted      i      fitted values for dependent variable
- 192  *          pval        p      p values
- 193  *          r2                 R Squared
- 194  *          resid       i      residuals
- 195  *          resvar             residual variance
- 196  *          rss                residual sum of squares
- 197  *          se          p      standard errors
- 198  *          sigma              standard error
- 199  *          tval        p      t values
- 200   
- 201   
- 205  *
- 214   
- 215  $EmbeddedCode Python:
- 216  import numpy as np
- 217  n2u = np.array(list(gams.get('n', keyType=KeyType.INT, valueFormat=ValueFormat.SKIP)),dtype=int)
- 218  u2n = np.zeros(n2u[-1]-n2u[0]+1, dtype=int)
- 219  for n,u in enumerate(n2u):
- 220    u2n[u-n2u[0]] = n+1
- 221  a = np.zeros((len(n2u),len(n2u)))  
- 222  for r in gams.get('Yl', keyType=KeyType.INT, keyFormat=KeyFormat.FLAT):
- 223    if r[0]>n2u[-1] or r[0]<n2u[0] or u2n[r[0]-n2u[0]] == 0: continue
- 224    if r[1]>n2u[-1] or r[1]<n2u[0] or u2n[r[1]-n2u[0]] == 0: continue
- 225    a[u2n[r[0]-n2u[0]]-1,u2n[r[1]-n2u[0]]-1] = r[2]
- 226  l = np.linalg.cholesky(a)
- 227  gams.set('L', [ (n2u[i],n2u[j],l[i,j]) for (i,j) in zip(*l.nonzero()) ], mapKeys=int)
- 228  $endEmbeddedCode L
-EXIT C:\Program Files (x86)\GAMS\42\inclib\linalg.gms
- 230   
- 231  SDPRelaxation = 0.5*sum(e, w(e)*(1 - Y.l(e)));
- 232   
- 234   
- 235  display SDPRelaxation;
- 236   
- 237  * Now do the random hyperplane r
- 238  Parameter r(n);
- 239   
- 240   
- 241  set hp      hyperplanes / hp1*hp10 /;
- 242  parameter rep_hp(hp,*)
- 243  loop(hp,
- 244     r(n) = uniform(-1,1);
- 245     S(n) = sum(i, L(n,i)*r(i)) < 0;
- 246     T(n) = yes;
- 247     T(S) =  no;
- 248     wS   = sum(e(S,T), w(S,T)) + sum(e(T,S), + w(T,S));
- 249     rep_hp(hp,'value of cut')              = ws;
- 250     rep_hp(hp,'value of best known bound') = SDPRelaxation;
- 251     rep_hp(hp,'relative gap')              = (SDPRelaxation-ws)/SDPRelaxation;
- 252     if(wS > maxwS, maxwS = wS; mingapS = (SDPRelaxation-ws)/SDPRelaxation; bestS(n) = S(n););
- 253  );
- 254  option clear=S;
- 255  S(bestS) = yes;
- 256  T(n)     = yes;
- 257  T(S)     =  no;
- 258   
- 259  display maxwS, mingapS, rep_hp;
- 260   
- 261  rep('value of cut')              = maxwS;
- 262  rep('value of best known bound') = SDPRelaxation;
- 263  rep('relative gap')              = (rep('value of best known bound')-rep('value of cut'))/rep('value of best known bound');
- 264   
- 265  put_utility 'log' / '### Value of cut:              ' rep('value of cut'):0:4;
- 266  put_utility 'log' / '### Value of best known bound: ' rep('value of best known bound'):0:4;
- 267  put_utility 'log' / '### Relative gap:              ' rep('relative gap'):0:16;
- 268   
- 269   
- 270  * use computed feasible solution as starting point for MIP solve
- 271  x.l(bestS)    = 1;
- 272  cut.l(e(i,j)) = x.l(i) xor x.l(j);
- 274   
- 276   
- 277   
- 279  * SCIP and COPT do this by default, for other solvers we need to enable it
- 282  option miqcp=cplex;
- 283  file cpxopt2 / cplex.op2 /;
- 284  putclose cpxopt2 'mipstart 1' /
- 285  'upperobjstop ' SDPRelaxation:0:16
- 286  putclose cpxopt2;
- 287  qubo.optFile = 2;
- 289  qubo.limrow   = 0;
- 290  qubo.limcol   = 0;
- 291  qubo.solprint = 0;
- 292  qubo.reslim   = 60-timeelapsed;
- 293  solve qubo max z using miqcp;
- 294   
- 295  option clear=S;
- 296  S(n) = x.l(n) > 0.5;
- 297  T(n)     = yes;
- 298  T(S)     =  no;
- 299   
- 300  parameter rep(*);
- 301  rep('value of cut')              = max(rep('value of cut'),z.l);
- 302  rep('value of best known bound') = min(rep('value of best known bound'),qubo.objest);
- 303  rep('relative gap')              = (rep('value of best known bound')-rep('value of cut'))/rep('value of best known bound');
- 305   
- 306  put_utility 'log' / '### Value of cut:              ' rep('value of cut'):0:4;
- 307  put_utility 'log' / '### Value of best known bound: ' rep('value of best known bound'):0:4;
- 308  put_utility 'log' / '### Relative gap:              ' rep('relative gap'):0:16;
- 309   
- 310  *write solution
- 313  embeddedCode Python:
- 314  import networkx as nx
- 315  # read input graph
- 316  g = nx.read_gml("G:/Studium/TVA/ProvideQ/toolbox-server/jobs/MAX_CUT/1/problem.gml", label=None)
- 317  # get cut value and bound ant turn them into graph attributes
- 318  attrs_g = {"cut_value": list(gams.get('rep'))[0][1], "bound": list(gams.get('rep'))[1][1]}
- 319  g.graph.update(attrs_g)
- 320  # get node partition
- 321  s = list(gams.get("S"))
- 322  t = list(gams.get("T"))
- 323  # turn node partitions into dicts
- 324  attr1 = {}
- 325  for n in s:
- 326      try:
- 327          attr1[int(n)] = {"Partition": 1}
- 328      except ValueError:
- 329          attr1[n] = {"Partition": 1}
- 330  attr2 = {}
- 331  for n in t:
- 332      try:
- 333          attr2[int(n)] = {"Partition": 2}
- 334      except ValueError:
- 335          attr2[n] = {"Partition": 2}
- 336  # make node partition a node attribute
- 337  nx.set_node_attributes(g, attr1)
- 338  nx.set_node_attributes(g, attr2)
- 339  lines = list(nx.generate_gml(g, stringizer=None))
- 340  with open("G:/Studium/TVA/ProvideQ/toolbox-server/jobs/MAX_CUT/1/problem_sol.gml", 'w') as fout:
- 341      if "Comment" in lines[1] or "id" in lines[2]:
- 342          for line in lines[:5]:
- 343              fout.write(line + '\n')
- 344      else:
- 345          for line in lines[:3]:
- 346              fout.write(line + '\n')
- 347              
- 348      for node in g.nodes(data=True):
- 349          if type(node[0]) == int:
- 350              strNode = "  node [\n    id " + f'{node[0]}' + "\n"
- 351          else:
- 352              strNode = "  node [\n    id " + "\"{}\"".format(node[0]) + "\n"
- 353          for key, val in node[-1].items():
- 354              if type(val) == int or type(val) == float:
- 355                  strNode += "    "+ str(key) + " " + f'{val}' + "\n"
- 356              else:
- 357                  strNode += "    "+ str(key) + " " + "\"{}\"".format(str(val)) + "\n"
- 358          strNode += "  ]"
- 359          fout.write(strNode + '\n')
- 360      for edge in g.edges(data=True):
- 361          if type(edge[0]) == int:
- 362              strEdge = "  edge [\n    source " + f'{edge[0]}' + "\n    target " + f'{edge[1]}' + "\n"
- 363          else:
- 364              strEdge = "  edge [\n    source " + "\"{}\"".format(edge[0]) + "\n    target " + "\"{}\"".format(edge[1]) + "\n"
- 365          for key, val in edge[-1].items():
- 366              if type(val) == int or type(val) == float:
- 367                  strEdge += "    " + str(key) + " " + f'{val}' + "\n"
- 368              else: 
- 369                  strEdge += "    " + str(key) + " " + "\"{}\"".format(str(val)) + "\n"
- 370          strEdge += "  ]"
- 371          fout.write(strEdge + '\n')
- 372      fout.write("]")
- 373  endembeddedCode
- 374   
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 2
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Include File Summary
-
-
-   SEQ   GLOBAL TYPE      PARENT   LOCAL  FILENAME
-
-     1        1 INPUT          0       0  G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\maxcut.gms
-     2      136 LIBINCLUDE     1     158  .C:\Program Files (x86)\GAMS\42\inclib\linalg.gms
-     3      229 EXIT           2     105  .C:\Program Files (x86)\GAMS\42\inclib\linalg.gms
-
-
-COMPILATION TIME     =        0.172 SECONDS      3 MB  42.0.0 3dcbfd09 WEX-WEI
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 3
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Range Statistics    SOLVE sdp Using LP From line 132
-
-
-RANGE STATISTICS (ABSOLUTE NON-ZERO FINITE VALUES)
-
-RHS       [min, max] : [        NA,        NA] - Zero values observed as well
-Bound     [min, max] : [ 1.000E+00, 1.000E+00]
-Matrix    [min, max] : [ 5.000E-01, 1.000E+00] - Zero values observed as well
-
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 4
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Model Statistics    SOLVE sdp Using LP From line 132
-
-
-MODEL STATISTICS
-
-BLOCKS OF EQUATIONS           1     SINGLE EQUATIONS            1
-BLOCKS OF VARIABLES           2     SINGLE VARIABLES            5
-NON ZERO ELEMENTS             5
-
-
-GENERATION TIME      =        0.000 SECONDS      4 MB  42.0.0 3dcbfd09 WEX-WEI
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 5
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Solution Report     SOLVE sdp Using LP From line 132
-
-
-               S O L V E      S U M M A R Y
-
-     MODEL   sdp                 OBJECTIVE  sdpobj
-     TYPE    LP                  DIRECTION  MINIMIZE
-     SOLVER  MOSEK               FROM LINE  132
-
-**** SOLVER STATUS     1 Normal Completion
-**** MODEL STATUS      1 Optimal
-**** OBJECTIVE VALUE               -1.0000
-
- RESOURCE USAGE, LIMIT          0.000        59.834
- ITERATION COUNT, LIMIT         3    2147483647
-*** This solver runs with a demo license. No commercial use.
-
-
-**** REPORT SUMMARY :        0     NONOPT
-                             0 INFEASIBLE
-                             0  UNBOUNDED
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 6
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-E x e c u t i o n
-
-
-----    235 PARAMETER SDPRelaxation        =        1.000  
-
-----    259 PARAMETER maxwS                =        1.000  best weight
-            PARAMETER mingapS              =  -7.7597E-11  min gap
-
-----    259 PARAMETER rep_hp  
-
-      value of ~  value of ~  relative ~
-
-hp1        1.000       1.000 -7.7597E-11
-hp2        1.000       1.000 -7.7597E-11
-hp3        1.000       1.000 -7.7597E-11
-hp4        1.000       1.000 -7.7597E-11
-hp5        1.000       1.000 -7.7597E-11
-hp6        1.000       1.000 -7.7597E-11
-hp7        1.000       1.000 -7.7597E-11
-hp8        1.000       1.000 -7.7597E-11
-hp9        1.000       1.000 -7.7597E-11
-hp10       1.000       1.000 -7.7597E-11
-
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 7
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Range Statistics    SOLVE qubo Using MIQCP From line 293
-
-
-RANGE STATISTICS (ABSOLUTE NON-ZERO FINITE VALUES)
-
-RHS       [min, max] : [        NA,        NA] - Zero values observed as well
-Bound     [min, max] : [ 1.000E+00, 1.000E+00] - Zero values observed as well
-Matrix    [min, max] : [ 1.000E+00, 2.000E+00]
-
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 8
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Model Statistics    SOLVE qubo Using MIQCP From line 293
-
-
-MODEL STATISTICS
-
-BLOCKS OF EQUATIONS           1     SINGLE EQUATIONS            1
-BLOCKS OF VARIABLES           2     SINGLE VARIABLES            3
-NON ZERO ELEMENTS             3     NON LINEAR N-Z              2
-CODE LENGTH                  13     CONSTANT POOL              16     DISCRETE VARIABLES          2
-
-
-GENERATION TIME      =        0.000 SECONDS      4 MB  42.0.0 3dcbfd09 WEX-WEI
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 9
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Solution Report     SOLVE qubo Using MIQCP From line 293
-
-
-               S O L V E      S U M M A R Y
-
-     MODEL   qubo                OBJECTIVE  z
-     TYPE    MIQCP               DIRECTION  MAXIMIZE
-     SOLVER  CPLEX               FROM LINE  293
-
-**** SOLVER STATUS     1 Normal Completion
-**** MODEL STATUS      1 Optimal
-**** OBJECTIVE VALUE                1.0000
-
- RESOURCE USAGE, LIMIT          0.000        59.742
- ITERATION COUNT, LIMIT         0    2147483647
- EVALUATION ERRORS             NA             0
---- GAMS/Cplex Link licensed for continuous and discrete problems.
-
-Reading parameter(s) from "G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\cplex.op2"
->>  mipstart 1
->>  upperobjstop 0.9999999999224030
-Finished reading from "G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\cplex.op2"
-
---- GMO Q Extraction (ThreePass): 0.00s
---- GMO setup time: 0.00s
---- GMO memory 0.50 Mb (peak 0.50 Mb)
---- Dictionary memory 0.00 Mb
---- Cplex 22.1.1.0 link memory 0.00 Mb (peak 0.00 Mb)
---- Starting Cplex
-
-
---- MIQP status (101): integer optimal solution.
---- Cplex Time: 0.00sec (det. 0.01 ticks)
-
---- Fixing integer variables and solving final QP...
-
-
---- Fixed MIQP status (1): optimal.
---- Cplex Time: 0.00sec (det. 0.00 ticks)
-
-
-Proven optimal solution
-MIP Solution:            1.000000    (0 iterations, 0 nodes)
-Final Solve:             1.000000    (0 iterations)
-
-Best possible:           1.000000
-Absolute gap:            0.000000
-Relative gap:            0.000000
-
-
-
-**** REPORT SUMMARY :        0     NONOPT
-                             0 INFEASIBLE
-                             0  UNBOUNDED
-                             0     ERRORS
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:32 Page 10
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-E x e c u t i o n
-
-
-**** REPORT FILE SUMMARY
-
- pyScript G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\225a\myEmb.dat
-cpxopt2 G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\cplex.op2
-
-
-EXECUTION TIME       =        0.141 SECONDS      4 MB  42.0.0 3dcbfd09 WEX-WEI
-
-
-USER: ProvideQ Project License                       S220922|0002CO-GEN
-      GAMS Software GmbH, Braunschweig Office                   DCE2398
-
-**** ************* BETA release
-**** GAMS Base Module 42.0.0 3dcbfd09 Jan 13, 2023   (BETA) WEI x86 64bit/MS Window
-**** ************* BETA release
-
-
-**** FILE SUMMARY
-
-Input      G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\maxcut.gms
-Output     G:\Studium\TVA\ProvideQ\toolbox-server\gams\maxCut\maxcut.lst
diff --git a/gams/maxcut/maxcut.lst b/gams/maxcut/maxcut.lst
deleted file mode 100644
index 99668abe..00000000
--- a/gams/maxcut/maxcut.lst
+++ /dev/null
@@ -1,427 +0,0 @@
-GAMS 40.3.0  f227c22a Sep 16, 2022          WEX-WEI x86 64bit/MS Windows - 02/02/23 23:10:23 Page 1
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-C o m p i l a t i o n
-
-
-   2   
-      Let G(N, E) denote a graph. A cut is a partition of the vertices N
-      into two sets S and T. Any edge (u,v) in E with u in S and v in T is
-      said to be crossing the cut and is a cut edge. The size of the cut is
-      defined to be sum of weights of the edges crossing the cut.
-       
-      This model presents a simple MIP formulation of the problem that is
-      seeded with a solution from the Goemans/Williamson randomized
-      approximation algorithm based on a semidefinite programming
-      relaxation. By default CSDP is used to solve the SDP.
-      Use --SDPSOLVER=MOSEK to switch to Mosek.
-       
-      The MaxCut instance tg20_7777 is available from the Biq Mac Library
-      and comes from applications in statistical physics.
-       
-       
-      Wiegele A., Biq Mac Library - Binary Quadratic and Max Cut Library.
-      http://biqmac.uni-klu.ac.at/biqmaclib.html
-       
-      Goemans M.X., and Williamson, D.P., Improved Approximation Algorithms
-      for Maximum Cut and Satisfiability Problems Using Semidefinite
-      Programming. Journal of the ACM 42 (1995), 1115-1145.
-      http://www-math.mit.edu/~goemans/PAPERS/maxcut-jacm.pdf
-       
-      Keywords: mixed integer linear programming, approximation algorithms,
-                convex optimization, randomized algorithms, maximum cut problem,
-                mathematics
-  31   
-  39   
-  40  Set n 'nodes';
-  41   
-  42  Alias (n,i,j);
-  43   
-  44  Parameter w(i,j) 'edge weights';
-  45   
-  46  Set e(i,j) 'edges';
-  47   
-  48  Set S(n), T(n), bestS(n);
-  49   
-  50  Scalar
-  51      wS      weight of cut S / -inf /
-  52      maxwS   best weight / -inf /
-  53      mingapS min gap / inf /;
-  54   
-  55  Scalar SDPRelaxation / inf /;
-  56   
-  57  parameter rep(*) / 'value of cut'              -inf
-  58                     'value of best known bound' +inf
-  59                     'relative gap'              NA   /;
-  60   
-  61   
-  67  $ offembeddedCode n w
-****                     $865
-**** 865  Problem in embedded code section
-  68   
-  69  * We want all edges to be i-j with i<j;
-  70  e(i,j)    = ord(i) < ord(j);
-****    $352,352      $141     $141
-**** 141  Symbol declared but no values have been assigned. Check for missing
-****         data definition, assignment, data loading or implicit assignment
-****         via a solve statement.
-****         A wild shot: You may have spurious commas in the explanatory
-****         text of a declaration. Check symbol reference list.
-**** 352  Set has not been initialized
-  71  w(e(i,j)) = w(i,j) + w(j,i);
-****      $352,352$141,352,352,352,352
-**** 141  Symbol declared but no values have been assigned. Check for missing
-****         data definition, assignment, data loading or implicit assignment
-****         via a solve statement.
-****         A wild shot: You may have spurious commas in the explanatory
-****         text of a declaration. Check symbol reference list.
-**** 352  Set has not been initialized
-  72  w(i,j)$(not e(i,j)) = 0;
-****    $352,352    $352,352
-**** 352  Set has not been initialized
-  73   
-  74  option e < w;
-  75  *option e:0:0:1, w:8:0:1; display n,e,w;
-  76   
-  77  * Simple MIP model
-  78  Variable
-  79     x(n)     'decides on what side of the cut'
-  80     cut(i,j) 'edge is in the cut'
-  81     z        'objective';
-  82   
-  83  Binary Variable x;
-  84   
-  85  Equation obj, xor1(i,j), xor2(i,j), xor3(i,j), xor4(i,j);
-  86   
-  87  obj..          z      =e= sum(e, w(e)*cut(e));
-  88   
-  89  xor1(e(i,j)).. cut(e) =l= x(i) + x(j);
-  90   
-  91  xor2(e(i,j)).. cut(e) =l= 2 - x(i) - x(j);
-  92   
-  93  xor3(e(i,j)).. cut(e) =g= x(i) - x(j);
-  94   
-  95  xor4(e(i,j)).. cut(e) =g= x(j) - x(i);
-  96   
-  97  Model maxcut / all /;
-  98   
-  99   
- 100  * QUBO Model
- 101  equation defQUBO;
- 102  defQUBO.. z =e= sum(e(i,j), w(e)*(sqr(x(i)) + sqr(x(j)) -2*x(i)*x(j)));
- 103  model qubo / defQUBO /;
- 104   
- 105   
-      Set up the SDP
-         max W*Y s.t. Y_ii = 1, Y positive semidefinite (psd)
-      We need to pass on the dual to csdp
-         min x1 + x2 + ... + xn s.t. X = F1*x1 + F2*x2 + ... + Fn*xn - W, X psd
-      with F_i = 1 for F_ii and 0 otherwise
- 114   
- 115   
- 116  Parameter L(i,j) 'Cholesky factor of Y';
- 117   
- 119   
- 120  Variable Y(i,j)    'PSDMATRIX';
- 121  Variable sdpobj    'objective function variable';
- 122  Equation sdpobjdef 'objective function W*Y';
- 123  sdpobjdef.. sum(e(i,j),w(i,j)*(Y(i,j)+Y(j,i))/2.0) + sum((i,j),eps*Y(i,j)) =E= sdpobj;
- 124  Y.fx(i,i) = 1.0;
-****       $352,352
-**** LINE    146 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 125  Model sdp / sdpobjdef /;
- 126   
- 127  option lp = MOSEK;
- 128  sdp.limrow   = 0;
- 129  sdp.limcol   = 0;
- 130  sdp.solprint = 0;
- 131  sdp.reslim   = 60-timeelapsed;
- 132  Solve sdp min sdpobj using lp;
-****                               $257
-**** LINE    154 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 257  Solve statement not checked because of previous errors
- 133   
- 134  Parameter Yl(i,j)   'level values of Y as parameter';
- 135  Yl(i,j) = Y.l(i,j);
-****     $352,352   $352,352
-**** LINE    157 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
-LIBINCLUDE C:\GAMS\40\inclib\linalg.gms
- 137  *
- 138  * LibInclude file to provide functionality of old matrix utilities cholesky via EmbeddedCode and numpy.linalg
- 139  * -c runs this at compile time, -e at execution time (default)
- 140  *
- 141  * Usage: $libInclude linalg [-c|-e] cholesky i A L
- 142  *        This calculates the Cholesky decomposition of a symmetric positive definite matrix A: A = LL^t
- 143  *        The matrix A is indexed over A(i,i)
- 144  *
- 145  * Usage: $libInclude linalg [-c|-e] eigenvalue i A AVal
- 146  *        This calculates the Eigenvalues of a symmetric positive definite matrix.
- 147  *        The matrix A is indexed over A(i,i). AVal(i) is indexed over i.
- 148  *
- 149  * Usage: $libInclude linalg [-c|-e] eigenvector i A AVal AVec
- 150  *        This calculates the Eigenvalues and Eigenvecors of a symmetric positive definite matrix.
- 151  *        The matrices A and AVec are indexed over (i,i). AVal(i) is indexed over i.
- 152  *
- 153  * Usage: $libInclude linalg [-c|-e] invert i A AInv
- 154  *        This calculates the inverse of a square matrix A: A*AInv = I
- 155  *        The matrices A and AInv are indexed over (i,i)
- 156  *
- 157  * Usage: $libInclude linalg [-c|-e] ols [-info=infoSym] [-cfival=val] [-intercept=0,1,2] [-CDFStudentT=fname] [-iCDFStudentT=fname] [-rcond=val] i p A y estimate
- 158  *        This estimates  the unknown parameters in a linear regression model.
- 159  *        The set i are the observations, the set p are the estimates. The matrices
- 160  *        A(i,p) contains the explanatory variable and y(i) the dependent variable.
- 161  *        On return the symbol estimate(p) will contain the estimated statistical
- 162  *        coefficients. The following parameters are available:
- 163  *          -intercept=0,1,2      For 0 no constant term or intercept will be added to
- 164  *                                the problem. For 1 a constant term will always be added.
- 165  *                                For 2 (default) the algorithm will add a constant term
- 166  *                                only if there is no data column with all ones in the matrix A.
- 167  *          -cfival=val           The confidence interval value. Default is 0.95. This impacts
- 168  *                                the calculation of confint_lo and confint_up.
- 169  *          -CDFStudentT=fname    The name of the cumulative distribution function of the
- 170  *                                StudentT distribution. The GAMS extrinisc function library
- 171  *                                stodclib provides this function. The default is CDFStudentT.
- 172  *                                The CDFStudentT is required to calculate pval.
- 173  *          -iCDFStudentT=fname   The name of the inverse cumulative distribution function of the
- 174  *                                StudentT distribution. The GAMS extrinisc function library
- 175  *                                stodclib provides this function. The default is iCDFStudentT.
- 176  *                                The iCDFStudentT is required to calculate confint_lo/up.
- 177  *          -rcond=val            Cut-off ratio for small singular values of A used as argument
- 178  *                                rcond to np.linalg.lstsq. Default is -1.
- 179  *          -mergetype=val        Determines if the symbols will be merged or replaced when loading
- 180  *                                into GAMS. Possible values are DEFAULT, REPLACE, and MERGE.
- 181  *                                Capitalization of this argument is important.
- 182  *                                See https://www.gams.com/latest/docs/UG_EmbeddedCode.html#UG_EmbeddedCode_Python
- 183  *        Additonal regression statistics can be requested via -info=infoSym. -info without the
- 184  *        infoSym, the name of info is used as a symbol name, e.g. -df sets GAMS scalar df.
- 185  *        The following statistics are available:
- 186  *          info        domain description
- 187  *          confint_lo  p      confidence interval (lower bound)
- 188  *          confint_up  p      confidence interval (upper bound)
- 189  *          covar       p,p    variance-covariance matrix
- 190  *          df                 degrees of freedom
- 191  *          fitted      i      fitted values for dependent variable
- 192  *          pval        p      p values
- 193  *          r2                 R Squared
- 194  *          resid       i      residuals
- 195  *          resvar             residual variance
- 196  *          rss                residual sum of squares
- 197  *          se          p      standard errors
- 198  *          sigma              standard error
- 199  *          tval        p      t values
- 200   
- 201   
- 205  *
- 214   
- 215  $EmbeddedCode Python:
- 216  import numpy as np
- 217  n2u = np.array(list(gams.get('n', keyType=KeyType.INT, valueFormat=ValueFormat.SKIP)),dtype=int)
- 218  u2n = np.zeros(n2u[-1]-n2u[0]+1, dtype=int)
- 219  for n,u in enumerate(n2u):
- 220    u2n[u-n2u[0]] = n+1
- 221  a = np.zeros((len(n2u),len(n2u)))  
- 222  for r in gams.get('Yl', keyType=KeyType.INT, keyFormat=KeyFormat.FLAT):
- 223    if r[0]>n2u[-1] or r[0]<n2u[0] or u2n[r[0]-n2u[0]] == 0: continue
- 224    if r[1]>n2u[-1] or r[1]<n2u[0] or u2n[r[1]-n2u[0]] == 0: continue
- 225    a[u2n[r[0]-n2u[0]]-1,u2n[r[1]-n2u[0]]-1] = r[2]
- 226  l = np.linalg.cholesky(a)
- 227  gams.set('L', [ (n2u[i],n2u[j],l[i,j]) for (i,j) in zip(*l.nonzero()) ], mapKeys=int)
- 228  $endEmbeddedCode L
-EXIT C:\GAMS\40\inclib\linalg.gms
- 230   
- 231  SDPRelaxation = 0.5*sum(e, w(e)*(1 - Y.l(e)));
- 232   
- 234   
- 235  display SDPRelaxation;
- 236   
- 237  * Now do the random hyperplane r
- 238  Parameter r(n);
- 239   
- 240   
- 241  set hp      hyperplanes / hp1*hp10 /;
- 242  parameter rep_hp(hp,*)
- 243  loop(hp,
- 244     r(n) = uniform(-1,1);
-****       $352
-**** LINE    173 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 245     S(n) = sum(i, L(n,i)*r(i)) < 0;
-****       $352     $352 $352,352,352
-**** LINE    174 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 246     T(n) = yes;
-****       $352
-**** LINE    175 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 247     T(S) =  no;
- 248     wS   = sum(e(S,T), w(S,T)) + sum(e(T,S), + w(T,S));
- 249     rep_hp(hp,'value of cut')              = ws;
- 250     rep_hp(hp,'value of best known bound') = SDPRelaxation;
- 251     rep_hp(hp,'relative gap')              = (SDPRelaxation-ws)/SDPRelaxation;
- 252     if(wS > maxwS, maxwS = wS; mingapS = (SDPRelaxation-ws)/SDPRelaxation; bestS(n) = S(n););
-****                                                                                  $352   $352
-**** LINE    181 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 253  );
- 254  option clear=S;
- 255  S(bestS) = yes;
- 256  T(n)     = yes;
-****    $352
-**** LINE    185 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 257  T(S)     =  no;
- 258   
- 259  display maxwS, mingapS, rep_hp;
- 260   
- 261  rep('value of cut')              = maxwS;
- 262  rep('value of best known bound') = SDPRelaxation;
- 263  rep('relative gap')              = (rep('value of best known bound')-rep('value of cut'))/rep('value of best known bound');
- 264   
- 265  put_utility 'log' / '### Value of cut:              ' rep('value of cut'):0:4;
- 266  put_utility 'log' / '### Value of best known bound: ' rep('value of best known bound'):0:4;
- 267  put_utility 'log' / '### Relative gap:              ' rep('relative gap'):0:16;
- 268   
- 269   
- 270  * use computed feasible solution as starting point for MIP solve
- 271  x.l(bestS)    = 1;
- 272  cut.l(e(i,j)) = x.l(i) xor x.l(j);
-****          $352,352    $352       $352
-**** LINE    201 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 274   
- 276   
- 277   
- 279  * SCIP and COPT do this by default, for other solvers we need to enable it
- 282  option miqcp=cplex;
- 283  file cpxopt2 / cplex.op2 /;
- 284  putclose cpxopt2 'mipstart 1' /
- 285  'upperobjstop ' SDPRelaxation:0:16
- 286  putclose cpxopt2;
- 287  qubo.optFile = 2;
- 289  qubo.limrow   = 0;
- 290  qubo.limcol   = 0;
- 291  qubo.solprint = 0;
- 292  qubo.reslim   = 60-timeelapsed;
- 293  solve qubo max z using miqcp;
-****                              $257
-**** LINE    266 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 257  Solve statement not checked because of previous errors
- 294   
- 295  option clear=S;
- 296  S(n) = x.l(n) > 0.5;
-****    $352     $352
-**** LINE    269 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 297  T(n)     = yes;
-****    $352
-**** LINE    270 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 352  Set has not been initialized
- 298  T(S)     =  no;
- 299   
- 300  parameter rep(*);
- 301  rep('value of cut')              = max(rep('value of cut'),z.l);
-****                                                               $141
-**** LINE    274 INPUT       C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-**** 141  Symbol declared but no values have been assigned. Check for missing
-****         data definition, assignment, data loading or implicit assignment
-****         via a solve statement.
-****         A wild shot: You may have spurious commas in the explanatory
-****         text of a declaration. Check symbol reference list.
- 302  rep('value of best known bound') = min(rep('value of best known bound'),qubo.objest);
- 303  rep('relative gap')              = (rep('value of best known bound')-rep('value of cut'))/rep('value of best known bound');
- 305   
- 306  put_utility 'log' / '### Value of cut:              ' rep('value of cut'):0:4;
- 307  put_utility 'log' / '### Value of best known bound: ' rep('value of best known bound'):0:4;
- 308  put_utility 'log' / '### Relative gap:              ' rep('relative gap'):0:16;
- 309   
- 310  *write solution
- 313  embeddedCode Python:
- 314  import networkx as nx
- 315  # read input graph
- 316  g = nx.read_gml("C:/Users/maxsc/Projects/toolbox-server/jobs/maxcut/2/problem.gml", label=None)
- 317  # get cut value and bound ant turn them into graph attributes
- 318  attrs_g = {"cut_value": list(gams.get('rep'))[0][1], "bound": list(gams.get('rep'))[1][1]}
- 319  g.graph.update(attrs_g)
- 320  # get node partition
- 321  s = list(gams.get("S"))
- 322  t = list(gams.get("T"))
- 323  # turn node partitions into dicts
- 324  attr1 = {}
- 325  for n in s:
- 326      try:
- 327          attr1[int(n)] = {"Partition": 1}
- 328      except ValueError:
- 329          attr1[n] = {"Partition": 1}
- 330  attr2 = {}
- 331  for n in t:
- 332      try:
- 333          attr2[int(n)] = {"Partition": 2}
- 334      except ValueError:
- 335          attr2[n] = {"Partition": 2}
- 336  # make node partition a node attribute
- 337  nx.set_node_attributes(g, attr1)
- 338  nx.set_node_attributes(g, attr2)
- 339  lines = list(nx.generate_gml(g, stringizer=None))
- 340  with open("C:/Users/maxsc/Projects/toolbox-server/jobs/maxcut/2/problem_sol.gml", 'w') as fout:
- 341      if "Comment" in lines[1] or "id" in lines[2]:
- 342          for line in lines[:5]:
- 343              fout.write(line + '\n')
- 344      else:
- 345          for line in lines[:3]:
- 346              fout.write(line + '\n')
- 347              
- 348      for node in g.nodes(data=True):
- 349          if type(node[0]) == int:
- 350              strNode = "  node [\n    id " + f'{node[0]}' + "\n"
- 351          else:
- 352              strNode = "  node [\n    id " + "\"{}\"".format(node[0]) + "\n"
- 353          for key, val in node[-1].items():
- 354              if type(val) == int or type(val) == float:
- 355                  strNode += "    "+ str(key) + " " + f'{val}' + "\n"
- 356              else:
- 357                  strNode += "    "+ str(key) + " " + "\"{}\"".format(str(val)) + "\n"
- 358          strNode += "  ]"
- 359          fout.write(strNode + '\n')
- 360      for edge in g.edges(data=True):
- 361          if type(edge[0]) == int:
- 362              strEdge = "  edge [\n    source " + f'{edge[0]}' + "\n    target " + f'{edge[1]}' + "\n"
- 363          else:
- 364              strEdge = "  edge [\n    source " + "\"{}\"".format(edge[0]) + "\n    target " + "\"{}\"".format(edge[1]) + "\n"
- 365          for key, val in edge[-1].items():
- 366              if type(val) == int or type(val) == float:
- 367                  strEdge += "    " + str(key) + " " + f'{val}' + "\n"
- 368              else: 
- 369                  strEdge += "    " + str(key) + " " + "\"{}\"".format(str(val)) + "\n"
- 370          strEdge += "  ]"
- 371          fout.write(strEdge + '\n')
- 372      fout.write("]")
- 373  endembeddedCode
- 374   
-
-**** 42 ERROR(S)   0 WARNING(S)
-GAMS 40.3.0  f227c22a Sep 16, 2022          WEX-WEI x86 64bit/MS Windows - 02/02/23 23:10:23 Page 2
-Goemans/Williamson Randomized Approximation Algorithm for MaxCut (MAXCUT,SEQ=338)
-Include File Summary
-
-
-   SEQ   GLOBAL TYPE      PARENT   LOCAL  FILENAME
-
-     1        1 INPUT          0       0  C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-     2      136 LIBINCLUDE     1     158  .C:\GAMS\40\inclib\linalg.gms
-     3      229 EXIT           2     105  .C:\GAMS\40\inclib\linalg.gms
-
-
-COMPILATION TIME     =        1.312 SECONDS      3 MB  40.3.0 f227c22a WEX-WEI
-
-
-USER: GAMS Demo license for Max Schweikart           G220922|0002CO-GEN
-      TVA at the Karlsruhe Institute of Technology, Germany    DL074179
-
-
-**** FILE SUMMARY
-
-Input      C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.gms
-Output     C:\Users\maxsc\Projects\toolbox-server\gams\maxcut\maxcut.lst
-
-**** USER ERROR(S) ENCOUNTERED
diff --git a/gams/sat/sat.lst b/gams/sat/sat.lst
deleted file mode 100644
index 09550aa8..00000000
--- a/gams/sat/sat.lst
+++ /dev/null
@@ -1,221 +0,0 @@
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 1
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-C o m p i l a t i o n
-
-
-   1  *input can be set through --CNFINPUT=<myfile.cnf>
-   2  *if no input is provided, use default input for demo
-   4   
-   5  set c             'clauses'
-   6      pn            'positive/negative' / '+', '-' /,
-   7      l             'literals'
-   8      cnf(c<,l,pn)  'clauses to literals mapping'
-   9  ;
-  10   
-  11  * read cnf file and load content into GAMS data structures
-  44   
-  45  * Simplest SAT Model
-  46  Binary variables
-  47      b(l)         'encodes yes (1) or no (0) for literal'
-  48  ;
-  49  Variable
-  50      obj          'dummy objective variable'
-  51  ;
-  52  Equation
-  53      defclause(c) 'define clauses'
-  54      defobj       'defined dummy objective'
-  55  ;
-  56   
-  57  defclause(c).. sum(cnf(c,l,'+'), b(l)) + sum(cnf(c,l,'-'), 1-b(l)) =g= 1;
-  58   
-  59  defobj..       obj =e= sum(l, b(l));
-  60   
-  61  model sat / all /;
-  62   
-  63  * set absolute termination criterion to a value satisfied by any feasible solution
-  64  sat.optca = card(l)+1;
-  65   
-  66  * solve SAT problem as MIP
-  67  solve sat max obj using mip;
-  68   
-  69  * write solution file
-  72  file fr / "G:\Studium\TVA\ProvideQ\toolbox-server\jobs\SAT\3\problem.sol" /; put fr;
-  73  put 'c Solution G:\Studium\TVA\ProvideQ\toolbox-server\jobs\SAT\3\problem.cnf';
-  74  if(sat.modelstat=1 or sat.modelstat=8,
-  75     put / 's cnf 1 ' card(l):0:0 ' ' card(c):0:0;
-  76     loop(l,
-  77       put$(b.l(l)<0.5) / 'v -' l.tl:0;
-  78       put$(b.l(l)>=0.5) / 'v ' l.tl:0;
-  79     );
-  80  elseif sat.solvestat=1,
-  81     put / 's cnf 0 ' card(l):0:0 ' ' card(c):0:0;
-  82  else
-  83     put / 's cnf -1 ' card(l):0:0 ' ' card(c):0:0;
-  84  );
-  85   
-
-
-COMPILATION TIME     =        0.062 SECONDS      3 MB  42.0.0 3dcbfd09 WEX-WEI
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 2
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-Equation Listing    SOLVE sat Using MIP From line 67
-
-
----- defclause  =G=  define clauses
-
-defclause(c0)..  b(1) =G= 1 ; (LHS = 0, INFES = 1 ****)
-     
-defclause(c1)..  b(2) =G= 1 ; (LHS = 0, INFES = 1 ****)
-     
-
----- defobj  =E=  defined dummy objective
-
-defobj..  - b(1) - b(2) + obj =E= 0 ; (LHS = 0)
-     
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 3
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-Column Listing      SOLVE sat Using MIP From line 67
-
-
----- b  encodes yes (1) or no (0) for literal
-
-b(1)
-                (.LO, .L, .UP, .M = 0, 0, 1, 0)
-        1       defclause(c0)
-       -1       defobj
-
-b(2)
-                (.LO, .L, .UP, .M = 0, 0, 1, 0)
-        1       defclause(c1)
-       -1       defobj
-
-
----- obj  dummy objective variable
-
-obj
-                (.LO, .L, .UP, .M = -INF, 0, +INF, 0)
-        1       defobj
-
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 4
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-Range Statistics    SOLVE sat Using MIP From line 67
-
-
-RANGE STATISTICS (ABSOLUTE NON-ZERO FINITE VALUES)
-
-RHS       [min, max] : [ 1.000E+00, 1.000E+00] - Zero values observed as well
-Bound     [min, max] : [ 1.000E+00, 1.000E+00] - Zero values observed as well
-Matrix    [min, max] : [ 1.000E+00, 1.000E+00]
-
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 5
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-Model Statistics    SOLVE sat Using MIP From line 67
-
-
-MODEL STATISTICS
-
-BLOCKS OF EQUATIONS           2     SINGLE EQUATIONS            3
-BLOCKS OF VARIABLES           2     SINGLE VARIABLES            3
-NON ZERO ELEMENTS             5     DISCRETE VARIABLES          2
-
-
-GENERATION TIME      =        0.000 SECONDS      4 MB  42.0.0 3dcbfd09 WEX-WEI
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 6
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-Solution Report     SOLVE sat Using MIP From line 67
-
-
-               S O L V E      S U M M A R Y
-
-     MODEL   sat                 OBJECTIVE  obj
-     TYPE    MIP                 DIRECTION  MAXIMIZE
-     SOLVER  CPLEX               FROM LINE  67
-
-**** SOLVER STATUS     1 Normal Completion
-**** MODEL STATUS      1 Optimal
-**** OBJECTIVE VALUE                2.0000
-
- RESOURCE USAGE, LIMIT          0.000 10000000000.000
- ITERATION COUNT, LIMIT         0    2147483647
---- GAMS/Cplex Link licensed for continuous and discrete problems.
---- GMO setup time: 0.00s
---- GMO memory 0.50 Mb (peak 0.50 Mb)
---- Dictionary memory 0.00 Mb
---- Cplex 22.1.1.0 link memory 0.00 Mb (peak 0.00 Mb)
---- Starting Cplex
-
-
---- MIP status (101): integer optimal solution.
---- Cplex Time: 0.00sec (det. 0.00 ticks)
-
---- Fixing integer variables and solving final LP...
-
-
---- Fixed MIP status (1): optimal.
---- Cplex Time: 0.00sec (det. 0.00 ticks)
-
-
-Proven optimal solution
-MIP Solution:            2.000000    (0 iterations, 0 nodes)
-Final Solve:             2.000000    (0 iterations)
-
-Best possible:           2.000000
-Absolute gap:            0.000000
-Relative gap:            0.000000
-
-
----- EQU defclause  define clauses
-
-          LOWER          LEVEL          UPPER         MARGINAL
-
-c0         1.0000         1.0000        +INF             .          
-c1         1.0000         1.0000        +INF             .          
-
-                           LOWER          LEVEL          UPPER         MARGINAL
-
----- EQU defobj              .              .              .             1.0000      
-
-  defobj  defined dummy objective
-
----- VAR b  encodes yes (1) or no (0) for literal
-
-         LOWER          LEVEL          UPPER         MARGINAL
-
-1          .             1.0000         1.0000         1.0000      
-2          .             1.0000         1.0000         1.0000      
-
-                           LOWER          LEVEL          UPPER         MARGINAL
-
----- VAR obj               -INF            2.0000        +INF             .          
-
-  obj  dummy objective variable
-
-
-**** REPORT SUMMARY :        0     NONOPT
-                             0 INFEASIBLE
-                             0  UNBOUNDED
-GAMS 42.0.0  3dcbfd09 Jan 13, 2023   (BETA) WEX-WEI x86 64bit/MS Windows - 02/10/23 11:10:48 Page 7
-G e n e r a l   A l g e b r a i c   M o d e l i n g   S y s t e m
-E x e c u t i o n
-
-
-**** REPORT FILE SUMMARY
-
-fr G:\Studium\TVA\ProvideQ\toolbox-server\jobs\SAT\3\problem.sol
-
-
-EXECUTION TIME       =        0.032 SECONDS      4 MB  42.0.0 3dcbfd09 WEX-WEI
-
-
-USER: ProvideQ Project License                       S220922|0002CO-GEN
-      GAMS Software GmbH, Braunschweig Office                   DCE2398
-
-**** ************* BETA release
-**** GAMS Base Module 42.0.0 3dcbfd09 Jan 13, 2023   (BETA) WEI x86 64bit/MS Window
-**** ************* BETA release
-
-
-**** FILE SUMMARY
-
-Input      G:\Studium\TVA\ProvideQ\toolbox-server\gams\sat\sat.gms
-Output     G:\Studium\TVA\ProvideQ\toolbox-server\gams\sat\sat.lst
diff --git a/gradlew b/gradlew
old mode 100644
new mode 100755