ERGO-Code
time_limit not work
With this mps file, HiGHS ignores time limit and does not stop at all. The logs are below.
$ ./highs problem.mps --time_limit 100
Running HiGHS 1.2.2 [date: 2022-08-25, git hash: d93ea2165]
Copyright (c) 2022 ERGO-Code under MIT licence terms
Number of BV entries in BOUNDS section is 2741
MIP problem has 7766 rows; 10391 cols; 84162 nonzeros; 2741 integer variables
Presolving model
3294 rows, 5866 cols, 54328 nonzeros
2335 rows, 4910 cols, 55633 nonzeros
2153 rows, 4694 cols, 31189 nonzeros
Solving MIP model with:
2153 rows
4694 cols (2697 binary, 0 integer, 1426 implied int., 571 continuous)
31189 nonzeros
Nodes | B&B Tree | Objective Bounds | Dynamic Constraints | Work
Proc. InQueue | Leaves Expl. | BestBound BestSol Gap | Cuts InLp Confl. | LpIters Time
0 0 0 0.00% -30718922.84515 inf inf 0 0 0 0 0.2s
R 0 0 0 0.00% -12820218.61013 687167347.6678 101.87% 0 0 0 1048 3.1s
Symmetry detection completed in 0.0s
Found 5 generators
B 0 0 0 0.00% -12819823.96427 -12809613.6905 0.08% 30 2 14 1147 7.1s
B 0 0 0 0.00% -12819823.96427 -12811608.88714 0.06% 34 2 25 1147 7.3s
Restarting search from the root node
Model after restart has 110 rows, 374 cols (316 bin., 0 int., 53 impl., 5 cont.), and 1525 nonzeros
4 0 0 0.00% -12818794.49298 -12811608.88714 0.06% 2 0 0 12963 8.8s
4 0 0 0.00% -12818794.49298 -12811608.88714 0.06% 2 2 3 12963 8.8s
So far it looks like your time limit is 100 seconds, but the time has only run for 8.8 seconds?
No, I mean that HiGHS does not stop after 100 seconds. It does not stop even after running for several hours. The logs only goes to 8.8 seconds, but HiGHs continues to run after that point.
Strange, to my knowledge @lgottwald is checking the time limit in the MIP solver. Perhaps it's in a part of the code where this isn't yet done
I will look where the solver spends the time. The time limit is checked but apparently the solver spents too much time in some routine between checks.
