[Cbc] relaxed infeasibilities

[James Howey]

I run a model and I get the following:

 

Cbc0010I After 22900 nodes, 4848 on tree, -220789.86 best solution, best
possible -520996.53 (246.43 seconds)

Cbc0003I Exiting on maximum nodes

Cbc0005I Partial search - best objective -220789.86 (best possible
-520996.53), took 715029 iterations and 23000 nodes (248.16 seconds)

Cbc0035I Maximum depth 233, 0 variables fixed on reduced cost

Cuts at root node changed objective from -522936 to -522936

Probing was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.018 seconds)

Gomory was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.001 seconds)

Knapsack was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.009 seconds)

Clique was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.000 seconds)

MixedIntegerRounding2 was tried 1 times and created 0 cuts of which 0 were
active after adding rounds of cuts (0.000 seconds)

FlowCover was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.010 seconds)

TwoMirCuts was tried 1 times and created 0 cuts of which 0 were active after
adding rounds of cuts (0.003 seconds)

Cgl0014I Postprocessing changed objective from -220789.86 to -512469.7 -
possible tolerance issue - try without preprocessing

2287 bounds tightened after postprocessing

 

883 relaxed row infeasibilities - summing to 8.19412e+09

694 relaxed row infeasibilities - summing to 8.29403e+09

794 relaxed row infeasibilities - summing to 2.69607e+09

 

Result - Stopped on node limit

 

Objective value:                -220789.85997650

Lower bound:                    -520996.529

Gap:                            0.58

Enumerated nodes:               23000

Total iterations:               715029

Time (CPU seconds):             259.72

Time (Wallclock seconds):       259.72

 

Total time (CPU seconds):       260.21   (Wallclock seconds):       260.21

 

The objective value reported is reasonable. But when I inspect the solution,
I get:

 

Stopped on iterations - objective value 59421894.99131566

**       0 E_worth(29)                       -59421894.991315655         0

 

It appears that the postprocessing is a little aggressive. And that the
summary printed doesn't reflect the postprocessing.

 

Is there a way I can tell the "relaxed row infeasibility" to back off and
report the summarized result?

 

I'll send models along if this is deemed worthy of investigation.

 

Thanks,

 

jkh

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20190127/d89cf1e1/attachment.html>