[Scip] Problem with SCIP_OBJSENSE_MAXIMIZE
Rostislav Stanek
rostislav.stanek at uni-graz.at
Sun Feb 15 17:57:42 CET 2015
Dear SCIP community,
I thank you for your response, but there was the following bug in my
program:
In order to solve the quadratic travelling salesman problem, I defined a
linearizatin similar to the (normal) travelling salesman problem first,
relaxed all subtour inequalities and solved the remaining quadratic
2-matching problem. Then I called “SCIPfreeTransform(sCIP);”, added some
subtour constraints and solved the enlarged ILP again. The problem was
that I set a lower bound before starting the new solving process. But I
have to set a upper bound in maximization problems. Clearly, this
problem does not occur by adapting the objective function, but does
occur by changing the optimization sense.
Thank you again and excuse this mistake of me.
Yours sincerely,
Rostislav Stanek
Dne 14.2.2015 v 17:07 dinesh kumar napsal(a):
> Dear,
> we would love to help you but we need to see the problem objective as
> well as your code for more information.
>
> Thank you
>
> On 13 February 2015 at 06:32, Rostislav Stanek
> <rostislav.stanek at uni-graz.at <mailto:rostislav.stanek at uni-graz.at>>
> wrote:
>
> Dear SCIP community,
>
> I have the following problem:
>
> By solving a minimization problem described as an ILP everything
> is working fine. Then I changed the signs by all objective
> function coefficients in order to obtain the corresponding
> maximization problem and everything is working fine as well. But
> by using the option
>
> SCIPsetObjsense(sCIP, SCIP_OBJSENSE_MAXIMIZE);
>
> an error occurs:
>
> src/lpi/lpi_cpx.c:3900: SCIPlpiGetNorms: Assertion
> `lpi->messagehdlr != ((void *)0)' failed.
>
> Program received signal SIGABRT, Aborted.
> 0x00007ffff694f0d5 in __GI_raise (sig=<optimized out>) at
> ../nptl/sysdeps/unix/sysv/linux/raise.c:64
> 64 ../nptl/sysdeps/unix/sysv/linux/raise.c: Adresář nebo soubor
> neexistuje.
> (gdb) bt
> #0 0x00007ffff694f0d5 in __GI_raise (sig=<optimized out>) at
> ../nptl/sysdeps/unix/sysv/linux/raise.c:64
> #1 0x00007ffff695283b in __GI_abort () at abort.c:91
> #2 0x00007ffff6947d9e in __assert_fail_base (fmt=<optimized out>,
> assertion=0x1711318 "lpi->messagehdlr != ((void *)0)",
> file=0x17104d8 "src/lpi/lpi_cpx.c", line=<optimized out>,
> function=<optimized out>) at assert.c:94
> #3 0x00007ffff6947e42 in __GI___assert_fail (assertion=0x1711318
> "lpi->messagehdlr != ((void *)0)", file=0x17104d8
> "src/lpi/lpi_cpx.c", line=3900,
> function=0x17116a0 "SCIPlpiGetNorms") at assert.c:103
> #4 0x0000000000ab6b73 in SCIPlpiGetNorms (lpi=0x22ccfd0,
> blkmem=0x20a4640, lpinorms=0x22da8a8) at src/lpi/lpi_cpx.c:3900
> #5 0x000000000054a087 in SCIPlpGetNorms (lp=0x22bf8b0,
> blkmem=0x20a4640, lpinorms=0x22da8a8) at src/scip/lp.c:12228
> #6 0x00000000006c18d9 in SCIPtreeStartProbing (tree=0x22da830,
> blkmem=0x20a4640, set=0x20a6660, lp=0x22bf8b0, strongbranching=0)
> at src/scip/tree.c:5954
> #7 0x0000000000656258 in SCIPstartProbing (scip=0x209d360) at
> src/scip/scip.c:29494
> #8 0x0000000000911c84 in performRandRounding (scip=0x209d360,
> heurdata=0x2201490, sol=0x2327910, cands=0x2b4dea0, ncands=62,
> propagate=1,
> result=0x7fffffffc818) at src/scip/heur_randrounding.c:94
> #9 0x0000000000912707 in performLPRandRounding (scip=0x209d360,
> heurdata=0x2201490, heurtiming=2, propagate=1, result=0x7fffffffc818)
> at src/scip/heur_randrounding.c:320
> #10 0x0000000000912d9c in heurExecRandrounding (scip=0x209d360,
> heur=0x2201520, heurtiming=2, nodeinfeasible=0, result=0x7fffffffc818)
> at src/scip/heur_randrounding.c:458
> #11 0x000000000051060a in SCIPheurExec (heur=0x2201520,
> set=0x20a6660, primal=0x22da7a0, depth=0, lpstateforkdepth=-1,
> heurtiming=2,
> nodeinfeasible=0, ndelayedheurs=0x7fffffffc81c,
> result=0x7fffffffc818) at src/scip/heur.c:454
> #12 0x0000000000699d8d in SCIPprimalHeuristics (set=0x20a6660,
> stat=0x209d450, prob=0x22db820, primal=0x22da7a0, tree=0x22da830,
> lp=0x22bf8b0,
> nextnode=0x0, heurtiming=2, nodeinfeasible=0,
> foundsol=0x7fffffffc9b0) at src/scip/solve.c:313
> #13 0x00000000006a0c82 in priceAndCutLoop (blkmem=0x20a4640,
> set=0x20a6660, messagehdlr=0x0, stat=0x209d450, transprob=0x22db820,
> origprob=0x209e020, primal=0x22da7a0, tree=0x22da830,
> lp=0x22bf8b0, pricestore=0x2301e70, sepastore=0x2389280,
> cutpool=0x231f270,
> delayedcutpool=0x2302030, branchcand=0x22c1e10,
> conflict=0x22daa20, eventfilter=0x22b6040, eventqueue=0x225a3c0,
> initiallpsolved=0,
> cutoff=0x7fffffffd02c, unbounded=0x7fffffffd030,
> lperror=0x7fffffffce54, pricingaborted=0x7fffffffce50) at
> src/scip/solve.c:2271
> #14 0x00000000006a3013 in solveNodeLP (blkmem=0x20a4640,
> set=0x20a6660, messagehdlr=0x0, stat=0x209d450,
> origprob=0x209e020, transprob=0x22db820,
> primal=0x22da7a0, tree=0x22da830, lp=0x22bf8b0,
> pricestore=0x2301e70, sepastore=0x2389280, cutpool=0x231f270,
> delayedcutpool=0x2302030,
> branchcand=0x22c1e10, conflict=0x22daa20,
> eventfilter=0x22b6040, eventqueue=0x225a3c0, initiallpsolved=0,
> newinitconss=0, cutoff=0x7fffffffd02c,
> unbounded=0x7fffffffd030, lperror=0x7fffffffce54,
> pricingaborted=0x7fffffffce50) at src/scip/solve.c:2788
> #15 0x00000000006a5513 in propAndSolve (blkmem=0x20a4640,
> set=0x20a6660, messagehdlr=0x0, stat=0x209d450,
> origprob=0x209e020, transprob=0x22db820,
> primal=0x22da7a0, tree=0x22da830, lp=0x22bf8b0,
> relaxation=0x22d9110, pricestore=0x2301e70, sepastore=0x2389280,
> branchcand=0x22c1e10,
> cutpool=0x231f270, delayedcutpool=0x2302030,
> conflict=0x22daa20, eventfilter=0x22b6040, eventqueue=0x225a3c0,
> focusnode=0x2306ff0, actdepth=0,
> timingmask=1, propagate=1, solvelp=1, solverelax=1,
> forcedlpsolve=0, nlperrors=0x7fffffffce30,
> fullpropagation=0x7fffffffce48,
> propagateagain=0x7fffffffce44, initiallpsolved=0x7fffffffce38,
> solvelpagain=0x7fffffffce40, solverelaxagain=0x7fffffffce3c,
> cutoff=0x7fffffffd02c, unbounded=0x7fffffffd030,
> lperror=0x7fffffffce54, pricingaborted=0x7fffffffce50,
> forcedenforcement=0x7fffffffce58)
> at src/scip/solve.c:3442
>
> Can you help me / explain it?
>
> Thank you for your response in advance.
>
> Yours sincerely,
> Rostislav Stanek
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de <mailto:Scip at zib.de>
> http://listserv.zib.de/mailman/listinfo/scip
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20150215/086a935e/attachment.html>
More information about the Scip
mailing list