search

ryanjoneil / python-zibopt

Automatically exported from code.google.com/p/python-zibopt
2 stars 1 forks source link

solver.restart() does not work #15

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
The FAQ states that solver.restart() is necessary to add additional 
constraints after a solution has been found for a second optimization. 

1. The example in the FAQ doesn't have a solver.restart() (and something 
like that doesn't work)

2. It doesn't work for me. No matter what I try, I get two different 
errors:

with restart():

3677 Segmentation fault

or without restart():

[src/scip/scip.c:9074] ERROR: invalid SCIP stage <8>
[src/scip/cons_linear.c:9891] ERROR: Error <0> in function call

Original issue reported on code.google.com by harald.schilly on 14 Feb 2010 at 3:28

GoogleCodeExporter commented 9 years ago 
Hi Harald,

SCIP is fussy about what state the solver is in when it takes particular 
actions, so
it's likely something was overlooked.  Two questions:

1. Can you run the TSP example against the US capitals data?  tsp.py is in the
examples directory and there json data file you want is in examples/tsp/.

2. Could you attach the model that is causing this error?

Thanks,
Ryan

Original comment by ryan%che...@gtempaccount.com on 14 Feb 2010 at 11:20

GoogleCodeExporter commented 9 years ago 
Hi, ad 1: yes, i can run both tsp examples. results look fine example output 
(No. 6 
out of 12)

[6] LENGTH: 696.0 / SUBTOURS: 2
    [0, 1, 41, 40, 39, 38, 37, 36, 35, 34, 33, 32, 31, 30, 29, 28, 27, 26, 25, 24, 
23, 10, 11, 9, 8, 7, 6, 5, 4, 3, 2]
    [12, 16, 21, 22, 20, 19, 18, 17, 15, 14, 13]

ad 2: i'm still modeling the pieces together and don't want to publish it here. 
once 
i've finished it i'll send it to you via email and explain where it brakes. 
i'll also 
look into the tsp example.

Original comment by harald.schilly on 15 Feb 2010 at 11:43

GoogleCodeExporter commented 9 years ago 
Is this still an issue?  If not, I will close it.

Original comment by ryan%che...@gtempaccount.com on 26 Jun 2010 at 5:15

GoogleCodeExporter commented 9 years ago 
Hi, I don't know. It turned out that scip 1.2.0 has a bug that hindered me 
using it. It always returned "infeasible" although there is a solution. Hence, 
for now, I've switched to gurobi. Next fall I'll need to calculate this problem 
again and then I'll try with scip. If the problem comes up again I'll provide a 
test case. Hence, for now you can close this ticket.

Original comment by harald.schilly on 26 Jun 2010 at 1:23

GoogleCodeExporter commented 9 years ago

Original comment by ryan%che...@gtempaccount.com on 27 Jun 2010 at 3:03