Decision Optimization

Decision Optimization

Delivers prescriptive analytics capabilities and decision intelligence to improve decision-making.


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  Cplex became very slow

    Posted 03/16/09 03:20 PM

    Originally posted by: SystemAdmin


    [shadi said:]

    hi,
    im using cplex to solve an ip problem, before some days the optimizer reached very quickly to the optimal solution (gap=0.00), now i dont know what was changed and the tool take a long long time to solve the same problem and it always write the following lines (the gap reduced very slow):


      29200 17556  1.54162e+007  183  1.54219e+007  1.53336e+007  4255304    0.57%
      29300 17615  1.53642e+007  172  1.54219e+007  1.53336e+007  4267140    0.57%
      29400 17671  1.53451e+007  185  1.54219e+007  1.53338e+007  4281620    0.57%
      29500 17728  1.54041e+007  151  1.54219e+007  1.53339e+007  4295952    0.57%
      29600 17777  1.53974e+007  145  1.54219e+007  1.53339e+007  4309637    0.57%
      29700 17826  1.53936e+007  166  1.54219e+007  1.53341e+007  4326397    0.57%
      29800 17874  1.54191e+007  131  1.54219e+007  1.53342e+007  4338426    0.57%
      29900 17925  1.53668e+007  169  1.54219e+007  1.53343e+007  4350650    0.57%
      30000 17989  1.53908e+007  178  1.54219e+007  1.53344e+007  4365292    0.57%
    Elapsed time = 2305.70 sec. (tree size = 246.36 MB, solutions = 17)
    Nodefile size = 118.74 MB (45.21 MB after compression)
      30100 18056  1.54107e+007  161  1.54219e+007  1.53345e+007  4376254    0.57%
      30200 18121  1.54200e+007  148  1.54219e+007  1.53345e+007  4388388    0.57%
      30300 18175  1.53741e+007  113  1.54219e+007  1.53347e+007  4402798    0.57%
      30400 18231        cutoff        1.54219e+007  1.53348e+007  4417975    0.57%
      30500 18287  1.54094e+007  164  1.54219e+007  1.53349e+007  4430055    0.56%
      30600 18346        cutoff        1.54219e+007  1.53349e+007  4444355    0.56%
      30700 18385  1.53779e+007  175  1.54219e+007  1.53351e+007  4460512    0.56%
      30800 18423        cutoff        1.54219e+007  1.53352e+007  4478865    0.56%
      30900 18481  1.54112e+007  182  1.54219e+007  1.53353e+007  4494268    0.56%
      31000 18521  1.53605e+007  177  1.54219e+007  1.53355e+007  4511206    0.56%
    Elapsed time = 2364.95 sec. (tree size = 253.63 MB, solutions = 17)
    Nodefile size = 125.72 MB (47.80 MB after compression)
      31100 18572        cutoff        1.54219e+007  1.53356e+007  4526829    0.56%
      31200 18625  1.53948e+007  177  1.54219e+007  1.53358e+007  4541257    0.56%
      31300 18685  1.53907e+007  162  1.54219e+007  1.53359e+007  4556465    0.56%
      31400 18743  1.54205e+007  130  1.54219e+007  1.53360e+007  4569653    0.56%
      31500 18790  1.54210e+007  160  1.54219e+007  1.53360e+007  4586345    0.56%
      31600 18829  1.53934e+007  172  1.54219e+007  1.53362e+007  4601212    0.56%
      31700 18892        cutoff        1.54219e+007  1.53362e+007  4612083    0.56%
      31800 18938        cutoff        1.54219e+007  1.53363e+007  4626522    0.56%
      31900 18984  1.53639e+007  178  1.54219e+007  1.53364e+007  4642026    0.55%
      32000 19030  1.53849e+007  159  1.54219e+007  1.53366e+007  4663728    0.55%
    Elapsed time = 2431.61 sec. (tree size = 260.58 MB, solutions = 17)
    Nodefile size = 132.71 MB (50.38 MB after compression)
      32100 19084  1.53963e+007  165  1.54219e+007  1.53367e+007  4680226    0.55%
      32200 19137        cutoff        1.54219e+007  1.53367e+007  4697114    0.55%
      32300 19196  1.54201e+007  170  1.54219e+007  1.53368e+007  4710410    0.55%
      32400 19250        cutoff        1.54219e+007  1.53369e+007  4723415    0.55%
      32500 19303  1.53809e+007  183  1.54219e+007  1.53371e+007  4739652    0.55%
      32600 19366  1.53724e+007  160  1.54219e+007  1.53372e+007  4752239    0.55%
      32700 19435  1.53820e+007  156  1.54219e+007  1.53373e+007  4762898    0.55%
      32800 19491        cutoff        1.54219e+007  1.53374e+007  4775648    0.55%
      32900 19531  1.54084e+007  181  1.54219e+007  1.53375e+007  4792352    0.55%
      33000 19583        cutoff        1.54219e+007  1.53375e+007  4805933    0.55%
    Elapsed time = 2495.53 sec. (tree size = 268.16 MB, solutions = 17)
    Nodefile size = 140.69 MB (53.29 MB after compression)
      33100 19633  1.54105e+007  140  1.54219e+007  1.53377e+007  4820741    0.55%
      33200 19688  1.53768e+007  164  1.54219e+007  1.53378e+007  4834569    0.55%
      33300 19740  1.53962e+007  163  1.54219e+007  1.53379e+007  4848859    0.54%
      33400 19792  1.53873e+007  160  1.54219e+007  1.53380e+007  4860936    0.54%
      33500 19852  1.53725e+007  178  1.54219e+007  1.53380e+007  4873458    0.54%
      33600 19910  1.53844e+007  189  1.54219e+007  1.53380e+007  4886703    0.54%
      33700 19971  1.54097e+007  166  1.54219e+007  1.53381e+007  4898033    0.54%
      33800 20024  1.54146e+007  170  1.54219e+007  1.53382e+007  4910659    0.54%
      33900 20066        cutoff        1.54219e+007  1.53382e+007  4925090    0.54%
      34000 20132  1.54032e+007  151  1.54219e+007  1.53383e+007  4936701    0.54%
    .
    .
    .


    how i can solve this problem??????
    thanks in advanced



    #DecisionOptimization
    #OPLusingCPLEXOptimizer


  • 2.  Re: Cplex became very slow

    Posted 03/16/09 06:10 PM

    Originally posted by: SystemAdmin


    [jfk said:]

    Hello,
    well, the following suggestion are probably just helping identify the issue, but first are you absolutely sure that you have changed nothing in the model? if the answer is yes, then:
    1. if you are trying to solve the model from the IDE have you tried to solve it with oplrun? oplrun gives the advantage not to have the IDE overhead (and leaves the convenience having the IDE)
    2. do you have CPLEX? if yes, you can dump the model into some .lp or .sav file and try to solve it with CPLEX interactive: simply try to find in the directory of CPLEX a file called cplex.exe; launch it; a dos-based tool will pop up; read your .lp or .sav file and solve it (there are a few commands, easy to follow).

    cheers

    #DecisionOptimization
    #OPLusingCPLEXOptimizer