[Symphony] solution export and var names
Joost
joost at dsj.nl
Sun Dec 18 15:07:07 EST 2011
FYI:
I ran into the same problem as described here:
http://list.coin-or.org/pipermail/symphony/2011-March/000592.html (using
Symphony 5.4.3). The same model solved within 30min to optimality with
Symphony 5.1.2
The reply on the basis condition number from Ted was a good reminder to
check the coefficients.
After some re-work I was able to elimate a couple of large coefficients
(in the order of 4e6/5e6).
Now it solves in 5.4.3 as well.
Could anyone help me with:
- Is there an example available the usage conditionNumber in
CoinUtils?
I don't know how to use that within the current Symphony code base.
- Or could Symphony give a warning in a log file via some option?
And what is a critical value for the condition number?
Could 10^6 still be ok?
"As a general rule of thumb, if the condition number κ(A) = 10^k, then
you may lose up to k digits of accuracy on top of what would be lost to
the numerical method due to loss of precision from arithmetic methods."
(http://en.wikipedia.org/wiki/Condition_number)
And finally, for what is worth, here is the log, it seems to get stuck
somewhere in the feasibility pump:
In iteration 4, before calling branch()
... no cuts were added.
Now displaying final relaxed solution...
Branching on variable openWhse
children: [49452342.352, 3,0] [49452342.352, 3,0]
Decided to dive...
*************************************************
* Now processing NODE 191 LEVEL 10
*************************************************
**** Starting iteration 5 ****
The LP value is: 51245073.298 [0,649]
Now displaying the relaxed solution ...
Row effectiveness: rownum: 15172 ineff: 6 deletable: 6
Removing deletable rows ...
Receiving/creating cuts...
Cuts in the local pool: 0
In iteration 5, before calling branch()
... no cuts were added.
Now displaying final relaxed solution...
feasibility pump: cuts discarded = 10
feasibility pump: starting iteration 0, time used = 0.00
fp: solve lp 0
feasibility pump: starting iteration 1, time used = 0.00
fp: solve lp 1
feasibility pump: starting iteration 2, time used = 0.00
fp: solve lp 2
feasibility pump: starting iteration 3, time used = 0.00
fp: solve lp 3
feasibility pump: starting iteration 4, time used = 0.00
fp: found solution with value = 48815888.342281
fp: gap = 0.908596
sp: solution pool size = 2
fp: found solution = 51688612.59 time = 0.00
leaving feasibility pump.
****** Found Better Feasible Solution !
****** After Calling Heuristics !
****** Cost: 51688612.589581
tighten_root_bounds: at node 191, tightening 0 bounds in root
tighten_root_bounds: at node 191, tightening 0 bounds in root
tighten_root_bounds: at node 191, tightening 0 bounds in root
tighten_root_bounds: at node 191, tightening 0 bounds in root
solution found before branching
Continue with this node.
**** Starting iteration 6 ****
The LP value is: 51245073.298 [0,0]
Now displaying the relaxed solution ...
Row effectiveness: rownum: 15166 ineff: 0 deletable: 0
Branching because of tailoff in value of objective function!
In iteration 6, before calling branch()
... no cuts were added.
Now displaying final relaxed solution...
feasibility pump: cuts discarded = 10
feasibility pump: starting iteration 0, time used = 0.00
fp: solve lp 0
feasibility pump: starting iteration 1, time used = 0.00
fp: solve lp 1
feasibility pump: starting iteration 2, time used = 0.00
fp: solve lp 2
feasibility pump: starting iteration 3, time used = 0.00
fp: solve lp 3
feasibility pump: starting iteration 4, time used = 0.00
fp: solve lp 4
feasibility pump: starting iteration 5, time used = 0.00
fp: solve lp 5
feasibility pump: starting iteration 6, time used = 0.00
fp: solve lp
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/symphony/attachments/20111218/ff5857b5/attachment-0001.html>
More information about the Symphony
mailing list