[Cmpl] Linearise nonlinear objective function (Indrajit Sen Gupta)

[Mike Steglich]

Thats correct ... one note from my side ... Cmpl is able to transform products of variables (if at  least one variable is integer) equivalently into a set of linear in-equations automatically. There is a subchapter in the manual that describes these functionalities.


Cheers,

Mike
> Am 09.03.2015 um 21:39 schrieb Peter Meldrum <peter.meldrum at googlemail.com>:
> 
> Indrajit:
> 
> Based on the information provided about the formulation, one simple option
> you could try is substituting with another variable to represent the
> nonlinear term:
> 
> Z[j,k] = x[j] * x[k]
> 
> The objective function will become linear in Z[j,k] and x[k]. Writing this
> in longer form for greater clarity:
> 
> Constant *x[k] - Constant * Z[j,k]
> 
> Then add simple constraints to enforce the needed relationship between
> Z[j,k] and x[j] and x[k].
> 
> I will leave this last part for you to do as an exercise. If you're not sure
> about the last part, please ask again.
> 
> Best regards,
> Peter
> 
> -----Original Message-----
> From: cmpl-bounces at list.coin-or.org [mailto:cmpl-bounces at list.coin-or.org]
> On Behalf Of cmpl-request at list.coin-or.org
> Sent: 09 March 2015 16:00
> To: cmpl at list.coin-or.org
> Subject: Cmpl Digest, Vol 31, Issue 4
> 
> Send Cmpl mailing list submissions to
> 	cmpl at list.coin-or.org
> 
> To subscribe or unsubscribe via the World Wide Web, visit
> 	http://list.coin-or.org/mailman/listinfo/cmpl
> or, via email, send a message with subject or body 'help' to
> 	cmpl-request at list.coin-or.org
> 
> You can reach the person managing the list at
> 	cmpl-owner at list.coin-or.org
> 
> When replying, please edit your Subject line so it is more specific than
> "Re: Contents of Cmpl digest..."
> 
> 
> Today's Topics:
> 
>   1. Linearizing a nonlinear objective function (Indrajit Sen Gupta)
>   2.  Benchmark / Statistics (Thorsten)
>   3. [CMPL] Readcsv error (Thorsten)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Mon, 9 Mar 2015 12:05:32 +0530
> From: Indrajit Sen Gupta <indrajitsg at gmail.com>
> To: "cmpl at list.coin-or.org" <cmpl at list.coin-or.org>
> Subject: [Cmpl] Linearizing a nonlinear objective function
> Message-ID:
> 	<CAB-BrmcAJ_sttf_-SthBwvuorvnbPht6J_j+pu6RWYAwhaEEMA at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
> 
> Hi All,
> 
> I had a basic question regarding linearizing non-linear objective functions.
> Currently the software that we have is running into issues as it does not
> support non-linear objective function along with binary decision variables.
> 
> One part of the objective function looks something like this:
> 
> constant * (1 - x[j]) * x[k]
> 
> where x's are binary variables (0/1). Is there a way to linearize the
> objective function in such situations?
> 
> Regards,
> Indrajit
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> <http://list.coin-or.org/pipermail/cmpl/attachments/20150309/1bb010b3/attach
> ment-0001.html>
> 
> ------------------------------
> 
> Message: 2
> Date: Mon, 9 Mar 2015 10:07:01 +0100
> From: Thorsten <thorten at hotmail.com>
> To: "cmpl at list.coin-or.org" <cmpl at list.coin-or.org>
> Subject: [Cmpl]  Benchmark / Statistics
> Message-ID: <DUB123-W434C9ABE063C9057693783B11B0 at phx.gbl>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> 
> Hi Mike,
> 
> thank you very much indeed for the quick reply! I scanned the appendix for
> CBC-Parameters but couldn't find a suitable one which enables me to get the
> required information. Even increasing the log(Level) to maximum doesn't
> provide the neccessary information. Do you possibly have a hint or idea in
> mind for a CBC-Command to get me the programs lower bound and the gap?
> Thanks for your support!
> 
> Best regards,
> Thorsten
> 
> Hi Thorsten,
> 
> You can find these information in the Cbc output. Cmpl does not provide
> functionalities to get it directly within a Cmpl model.
> 
> Cheers,
> 
> Mike
> 
> 
> 
>> Am 06.03.2015 um 16:57 schrieb Thorsten <thorten at hotmail.com>:
>> 
>> Dear CMPL-Community,
>> 
>> I have implemented a capacitated vehicle routing problem with time windows
> in CMPL and it runs within coliop. In my thesis I want to compare the
> performance of CMPL and the CBC-solver with the performance of Fico Xpress.
> For the numerical analysis I will need some more statistics than I am able
> to get so far. Besides the objective value and the running time of the model
> I will need the Lower bound of the program and the gap. Can you probably
> help me how I can have these values printed out?
>> Thank you very much in advance,
>> Thorsten 		 	   		  
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> <http://list.coin-or.org/pipermail/cmpl/attachments/20150309/25390df5/attach
> ment-0001.html>
> 
> ------------------------------
> 
> Message: 3
> Date: Mon, 9 Mar 2015 10:12:18 +0100
> From: Thorsten <thorten at hotmail.com>
> To: "cmpl at list.coin-or.org" <cmpl at list.coin-or.org>
> Subject: [Cmpl] [CMPL] Readcsv error
> Message-ID: <DUB123-W19BD131338FFFD8FDBF555B11B0 at phx.gbl>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> Hi folks!
> 
> I am trying to read parameters from a csv-file and have therefore
> implemented the following code:
> serv[]:=        readcsv("Dist5C.csv");
> 
> Unfortunately I get the following error message:
> error (input/output): file Minnertry5C.cmpl line 18: Input file 'Dist5C.csv'
> not found
> 
> The csv-file is in the same directory as the cmpl-file.
> Does anybody have an idea what the problem could be?
> 
> Best regards,
> Thorsten
> 
> 		 	   		  
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> <http://list.coin-or.org/pipermail/cmpl/attachments/20150309/6be7d5d2/attach
> ment-0001.html>
> 
> ------------------------------
> 
> _______________________________________________
> Cmpl mailing list
> Cmpl at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cmpl
> 
> End of Cmpl Digest, Vol 31, Issue 4
> ***********************************
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.5646 / Virus Database: 4299/9259 - Release Date: 03/09/15
> 
> _______________________________________________
> Cmpl mailing list
> Cmpl at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cmpl