[Help-glpk] MathProg (GMPL) with Tables (i.e. SQLIte3) and CBC MILP solver works

[Noli Sicad]

Hi,

CBC 2.7.7 and GLPK-4.45 works (compile and install) out of the box in Mac OS X.

See this log (below).

It is quick, it takes only 9 minutes. Unlike the freemps format solves
in CBC as well, takes 17 minutes to solve this MIP model.

The CBC MathProg - ThridParty is ported by Stefan Vigerske who works
with GAMS now.

http://www.math.hu-berlin.de/~stefan/

The CBC MathProg does not return the solution to MathProg environment,
just input to CBC solver. It is only working using GLPK-4.45.

Noli

####

Nolis-MacBook-Pro:Case_Studies nsicad$ cbc
MIP_model_I_P6r_spl_1500_working_evenflow_0025p.mod% -printi all
-solve -solu solution.out
Welcome to the CBC MILP Solver
Version: 2.7.7
Build Date: Nov 20 2012

command line - cbc
MIP_model_I_P6r_spl_1500_working_evenflow_0025p.mod% -printi all
-solve -solu solution.out (default strategy 1)
GMPL model file ./MIP_model_I_P6r_spl_1500_working_evenflow_0025p.mod
and data file ./
Reading model section from
./MIP_model_I_P6r_spl_1500_working_evenflow_0025p.mod...
Reading data section from
./MIP_model_I_P6r_spl_1500_working_evenflow_0025p.mod...
599 lines were read
Reading Origin1...
Connected to SQLite 3.7.6.2 -
/Users/nsicad/Documents/A_Document_1/Pete_book/Lincoln_stands.sqlite
SELECT ID FROM Lincoln_stands
Display statement at line 41
STAND:
   1
   2
   3
   4
   5
   6
..

FlowCover was tried 100 times and created 0 cuts of which 0 were
active after adding rounds of cuts (0.061 seconds)
TwoMirCuts was tried 100 times and created 87 cuts of which 0 were
active after adding rounds of cuts (0.131 seconds)

Result - Optimal solution found

Objective value:                -345556.66666667
Enumerated nodes:               1007292
Total iterations:               12059944
Time (CPU seconds):             543.33
Time (Wallclock seconds):       546.31

Total time (CPU seconds):       543.37   (Wallclock seconds):       546.37

Nolis-MacBook-Pro:Case_Studies nsicad$