CODEWITHSUNDEEP
juliajulia-jumpcoin-or-cbccoin-or
RobIris
RobIris

Reputation: 21

Julia JuMP Cbc solver hangs infinitely without message and without exiting

I'm using Julia language (Version 1.3.1), JuMP package (Version 0.20.1) and Cbc package (Version 0.6.6) to solve an optimization problem in a docker container with ubuntu:16.04. The optimizer Cbc seems to be hung, with 100% cpu usage, without exiting and without any message. The problems happens rarely on similar problem and seems to be not replicable: if I run the same code with the same data it doesn't hang anymore. Hope that backtrace got through gdb can be useful.

I can share my model, if needed. It has 11520 variables, 4652 constraints, 10080 variables used in linear objective function.

This is the log of Cbc optimizer:

Welcome to the CBC MILP Solver Version: 2.10.3 Build Date: Oct 7 2019

command line - Cbc_C_Interface -threads 0 -seconds 360.0 -maxNodes 30000 -logLevel 1 -solve -quit (default strategy 1) seconds was changed from 1e+100 to 360 maxNodes was changed from 2147483647 to 30000 Continuous objective value is 2.3607e+08 - 0.11 seconds Cgl0002I 3197 variables fixed Cgl0005I 7 SOS with 8323 members Cgl0004I processed model has 15 rows, 8323 columns (8323 integer (8323 of which binary)) and 26556 elements Cbc0045I Fixing only non-zero variables. Cbc0045I Warning: mipstart values could not be used to build a solution.

Here Cbc seems to be hung and becomes unresponsive, with 100% cpu usage.

Here the backtrace on the running pid process:

#0 0x00007f163c3facc9 in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #1 0x00007f163c4125b3 in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #2 0x00007f163c467586 in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #3 0x00007f163c46aebc in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #4 0x00007f163c40594a in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #5 0x00007f163c29afbe in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #6 0x00007f163c2ad844 in ?? () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #7 0x00007f163b8ea31f in CbcHeuristicDive::solution(double&, int&, int&, OsiRowCut**, CbcSubProblem&, double*) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbc.so.3 #8 0x00007f163b8ebf42 in CbcHeuristicDive::solution(double&, double*) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbc.so.3 #9 0x00007f163b938fd2 in CbcModel::solveWithCuts(OsiCuts&, int, CbcNode*) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbc.so.3 #10 0x00007f163b9472d7 in CbcModel::branchAndBound(int) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbc.so.3 #11 0x00007f163c214c47 in CbcMain1(int, char const, CbcModel&, int ()(CbcModel, int), CbcSolverUsefulData&) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #12 0x00007f163c2252ae in CbcMain1(int, char const**, CbcModel&) () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #13 0x00007f163c19bc50 in Cbc_solve () from target:/root/.julia/packages/Cbc/vWzyC/deps/usr/lib/libCbcSolver.so #14 0x00007f16698e7e71 in ?? () #15 0x000000000000000c in ?? () #16 0x00007fff70694480 in ?? () #17 0x00007f16604ce110 in ?? () #18 0x000000000000262e in ?? () #19 0x0000000000000006 in ?? () #20 0x00007fff70694480 in ?? () #21 0x00007f165966ab40 in ?? () #22 0x00007f164a7ce1d0 in ?? () #23 0x00007f164a7ce220 in ?? () #24 0x00007f164a7ce1d0 in ?? () #25 0x00007f1688be7b00 in ?? () at /buildworker/worker/package_linux64/build/src/array.c:738 from target:/opt/julia/bin/../lib/libjulia.so.1 #26 0x00007f163d909af0 in ?? () #27 0x00007f164439d3c0 in ?? () #28 0x00007f1689524200 in ?? () #29 0x0000000000000000 in ?? ()

Using next command in gdb console, than a StackOverflowError() error is catched on CbC.

Has the objective function too many terms?

Any help is really appreciable.

Thank you

Upvotes: 1

Views: 668

Answers (2)

MrBurch
MrBurch

Reputation: 1

You can set the time limit using the seconds parameter as follows.

For newer package versions:

model = Model(optimizer_with_attributes(Cbc.Optimizer
                                        ,"seconds" => 60
                                        ,"threads" => 4
                                        ,"loglevel" => 0
                                        ,"ratioGap" => 0.0001))

Or like this for older package versions:

model = Model(with_optimizer(Cbc.Optimizer
                             ,seconds=60
                             ,threads=4
                             ,loglevel=0
                             ,ratioGap=0.0001))

Upvotes: 0

Oscar Dowson
Oscar Dowson

Reputation: 2574

This appears to be an issue with Cbc. It's impossible to provide more advice without a reproducible example. I suggest you try to simplify your model and create an MPS file.

Upvotes: 0

Related Questions