Gurobi optimization problem

77 views
Skip to first unread message

XM XM

unread,
Jul 12, 2017, 10:38:50 PM7/12/17
to YALMIP
Dear Prof. Lofberg,

I have an optimization problem. It seems that this question is not directly related to YALMIP, but your comments/suggestions are invaluable.

I use YALMIP with Gurobi and try to work out a Quadratic Programming problem. It runs infinitely. A screen shot of the output is as follows. What could be the possible problem?

Thanks again!

Warning for adding variables: zero or small (< 1e-13) coefficients, ignored
Optimize a model with 195832 rows, 115240 columns and 1940862 nonzeros
Model has 57600 quadratic objective terms
Coefficient statistics:
  Matrix range    [1e-10, 1e+02]
  Objective range [0e+00, 0e+00]
  Bounds range    [0e+00, 0e+00]
  RHS range       [2e-10, 1e+00]
Presolve removed 124306 rows and 6546 columns
Presolve time: 1.00s
Presolved: 71526 rows, 108694 columns, 1444167 nonzeros
Presolved model has 57600 quadratic objective terms
Ordering time: 0.90s

Barrier statistics:
 Dense cols : 459
 Free vars  : 4640
 AA' NZ     : 3.099e+06
 Factor NZ  : 5.003e+06 (roughly 110 MBytes of memory)
 Factor Ops : 1.099e+09 (less than 1 second per iteration)
 Threads    : 2

                  Objective                Residual
Iter       Primal          Dual         Primal    Dual     Compl     Time
   0   1.58072560e+09 -8.84121449e+09  1.00e+03 1.00e+03  9.43e+05     3s
   1   3.38314775e+08 -5.86209628e+09  6.17e+02 5.41e+02  5.66e+05     3s
   2   1.93925391e+09 -5.31291900e+09  1.12e+02 7.61e+01  1.32e+05     3s
   3   5.71984753e+08 -2.29761075e+09  3.17e+01 2.15e+01  4.00e+04     4s
   4   1.30849027e+08 -1.03957791e+09  5.93e+00 2.98e+00  1.07e+04     4s
   5   1.60806546e+07 -4.66274013e+08  8.41e-01 5.66e-01  3.58e+03     4s
   6   3.25849307e+06 -1.47817422e+08  6.20e-02 6.20e-02  9.28e+02     5s

..................

 91773   1.02775880e+02  1.02774802e+02  6.33e-09 2.23e-10  6.13e-09  52505s
 91774   1.02775880e+02  1.02774802e+02  6.42e-09 2.23e-10  6.13e-09  52506s
 91775   1.02775880e+02  1.02774802e+02  6.42e-09 2.23e-10  6.13e-09  52507s
 91776   1.02775880e+02  1.02774802e+02  6.42e-09 2.23e-10  6.13e-09  52507s
 91777   1.02775880e+02  1.02774802e+02  6.52e-09 2.23e-10  6.13e-09  52508s
 91778   1.02775880e+02  1.02774803e+02  6.51e-09 2.23e-10  6.13e-09  52508s
 91779   1.02775880e+02  1.02774803e+02  6.51e-09 2.22e-10  6.13e-09  52509s
 91780   1.02775880e+02  1.02774803e+02  6.50e-09 2.22e-10  6.13e-09  52509s
 91781   1.02775880e+02  1.02774803e+02  6.50e-09 2.23e-10  6.13e-09  52510s
 91782   1.02775880e+02  1.02774803e+02  6.52e-09 2.22e-10  6.13e-09  52511s
 91783   1.02775880e+02  1.02774803e+02  6.52e-09 2.22e-10  6.13e-09  52511s
 91784   1.02775880e+02  1.02774803e+02  6.52e-09 2.23e-10  6.13e-09  52512s
 91785   1.02775880e+02  1.02774803e+02  6.52e-09 2.23e-10  6.13e-09  52513s
 91786   1.02775880e+02  1.02774803e+02  6.52e-09 2.23e-10  6.13e-09  52514s
 91787   1.02775880e+02  1.02774803e+02  6.52e-09 2.22e-10  6.13e-09  52514s
 91788   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52516s
 91789   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52516s
 91790   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52517s
 91791   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52518s
 91792   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52519s
 91793   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52520s
 91794   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52521s
 91795   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52522s
 91796   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52522s
 91797   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52523s
 91798   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52523s
 91799   1.02775880e+02  1.02774803e+02  6.69e-09 2.22e-10  6.13e-09  52524s
 91800   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52524s
 91801   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52525s
 91802   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52526s
 91803   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52527s
 91804   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52528s
 91805   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52529s
 91806   1.02775880e+02  1.02774803e+02  6.60e-09 2.22e-10  6.13e-09  52530s

Johan Löfberg

unread,
Jul 13, 2017, 2:52:28 AM7/13/17
to YALMIP
Try tweaking some solver options.

XM XM

unread,
Jul 13, 2017, 11:37:32 AM7/13/17
to YALMIP
Do you mean YALMIP options, like MaxIter, or Gurobi's own tuning parameters?

Thanks

Johan Löfberg

unread,
Jul 13, 2017, 11:44:25 AM7/13/17
to YALMIP
All solver related stuff is done through solver specific options

>> ops = sdpsettings;ops.gurobi

ans = 

  struct with fields:

         BarIterLimit: Inf
               Cutoff: Inf
       IterationLimit: Inf
            NodeLimit: Inf
        SolutionLimit: Inf
            TimeLimit: Inf
           BarConvTol: 1.0000e-08
        BarQCPConvTol: 1.0000e-06
       FeasibilityTol: 1.0000e-06
           IntFeasTol: 1.0000e-06
         MarkowitzTol: 0.0078
               MIPGap: 1.0000e-04
            MIPGapAbs: 1.0000e-10
        OptimalityTol: 1.0000e-06
               PSDTol: 1.0000e-06
          InfUnbdInfo: 0
           NormAdjust: -1
             ObjScale: 0
         PerturbValue: 2.0000e-04
                 Quad: -1
            ScaleFlag: 1
              Sifting: -1
           SiftMethod: -1
       SimplexPricing: -1
        BarCorrectors: -1
       BarHomogeneous: -1
             BarOrder: -1
            Crossover: -1
       CrossoverBasis: 0
              QCPDual: 0
            BranchDir: 0
        ConcurrentMIP: 1
           Heuristics: 0.0500
      ImproveStartGap: 0
    ImproveStartNodes: Inf
     ImproveStartTime: Inf
          MinRelNodes: 0
             MIPFocus: 0
          MIQCPMethod: -1
          NodefileDir: '.'
        NodefileStart: Inf
           NodeMethod: 1
           PumpPasses: 0
                 RINS: -1
       SolutionNumber: 0
          SubMIPNodes: 500
             Symmetry: -1
            VarBranch: -1
         ZeroObjNodes: 0
           TuneOutput: 2
          TuneResults: -1
        TuneTimeLimit: -1
           TuneTrials: 2
                 Cuts: -1
           CliqueCuts: -1
            CoverCuts: -1
        FlowCoverCuts: -1
         FlowPathCuts: -1
         GUBCoverCuts: -1
          ImpliedCuts: -1
           MIPSepCuts: -1
              MIRCuts: -1
             ModKCuts: -1
          NetworkCuts: -1
           SubMIPCuts: -1
         ZeroHalfCuts: -1
         CutAggPasses: -1
            CutPasses: -1
         GomoryPasses: -1
              AggFill: 10
            Aggregate: 1
      DisplayInterval: 5
       DualReductions: 1
        FeasRelaxBigM: 1000000
            IISMethod: -1
              LogFile: ''
               Method: -1
         NumericFocus: 0
             PreCrush: 0
            PreDepRow: -1
              PreDual: -1
        PreQLinearize: -1
            PrePasses: -1
             Presolve: -1
          PreSparsify: 0
           ResultFile: ''
              Threads: 0


XM XM

unread,
Jul 13, 2017, 12:04:13 PM7/13/17
to YALMIP
Hi Prof. Lofberg,

Thanks! Here is the list of the Gurobi tunable parameters with instructions. Post it here for other people who are interested.

http://www.gurobi.com/documentation/6.0/refman/parameters.html
Reply all
Reply to author
Forward
0 new messages