[Cmpl] Problem when sets contain numeric values as strings and 'long line name' situation
Mike Steglich
mike.steglich at th-wildau.de
Wed May 25 08:56:37 EDT 2016
Hi Maria,
Thank you for the information that the line name issue is due to an MPS export and not a bug.
I created a bug fix of the other problem you mentioned. You will receive a link to my dropbox with an fixed beta version until the end of the day. It would be great if you could take a look at it.
A think (or hope) that is the last necessary fix before I will release Cmpl v1.11 in the next couple of days.
Thanks,
Mike
> Am 25.05.2016 um 14:30 schrieb Maria Giatsoglou <maria.giatsoglou at gmail.com>:
>
> Hi Mike,
>
> Thank you for the prompt reply.
> I will apply a workaround for the string values containing only numbers, but I will be looking forward to a new release that will solve this issue since currently it is rather cumbersome to work with such values.
>
> Indeed, the second problem I described is due to an MPS export. After removing this configuration the issue disappeared.
>
> All the best,
> Maria
>
>
> On Tue, May 24, 2016 at 7:17 PM, Mike Steglich <mike.steglich at th-wildau.de <mailto:mike.steglich at th-wildau.de>> wrote:
> Hi Maria,
>
> Sorry … there actually a simple bug. If in a CmplData file an entry enclosed in double quotes only contains numbers or looks like a float then it is interpreted as an integer or a real number. I will change it as soon as possible.
>
> The maximum length of a line name is 100. Only for the export to a MPS file it is restricted to nine characters. Could you send your jCmpl programme (directly to me)?
>
> Thanks,
>
> Mike
>
>> Am 24.05.2016 um 17:45 schrieb Maria Giatsoglou <maria.giatsoglou at gmail.com <mailto:maria.giatsoglou at gmail.com>>:
>>
>> Hi Mike,
>>
>> Please find attached the model I use and the cdat file.
>> I noticed that the first error ("Implicit type conversion from string to integer is not possible") occurs both when using the Coliop environment and the Cmpl Java API.
>> If I change the numeric ids into string ones e.g. by appending a letter before each id in the 'items' set, then the model can be solved without problems in Coliop.
>> By using jCmpl, however, I get the "Line name 'fm[1220],[1],[2]' is too long" error.
>> The attached cdat file is the one produced by jCmpl, which is the main way I use Cmpl (I use Coliop only for testing).
>>
>> I am looking forward for your response.
>>
>> Thank you once more for your help!
>>
>> Maria
>>
>> On Mon, May 23, 2016 at 10:31 PM, Mike Steglich <mike.steglich at th-wildau.de <mailto:mike.steglich at th-wildau.de>> wrote:
>> Hello Maria,
>>
>> can you send your model directly to me?
>>
>> Thanks,
>>
>> Mike
>>
>> Am 23.05.2016 um 11:36 schrieb Maria Giatsoglou <maria.giatsoglou at gmail.com <mailto:maria.giatsoglou at gmail.com>>:
>>
>>> Hi Mike,
>>>
>>> I have encountered the following situation with CMPL:
>>>
>>> In my model I am importing the following parameters set from a cdat file
>>>
>>> %M set[3] <
>>> "7" "2" "10"
>>> "1" "12" "7"
>>> >
>>>
>>> Then, I use this set in the following expression
>>>
>>> flag := 0;
>>> it {[i,j,k] in M, i != "off", j != "off", k != "off":
>>> flag:= 1; break;
>>> }
>>>
>>> By doing this I get the following error: "Implicit type conversion from string to integer is not possible".
>>> I thought that by using double quotes, the values of set M would be interpreted and treated as strings in CMPL (as was my intention), but I found out that using
>>>
>>> %M set[3] <
>>> "7" "2" "10"
>>> "1" "12" "7"
>>> >
>>>
>>> or
>>>
>>> %M set[3] <
>>> 7 2 10
>>> 1 12 7
>>> >
>>>
>>> makes no difference.
>>>
>>> Is there a way to circumvent this issue?
>>>
>>>
>>> Additionally, when I try to solve my model I get the following error message:
>>> "Line name 'fm[3835],[1],[2]' is too long"
>>> Can I somehow change the maximum line name length in CMPL?
>>>
>>> Thanks!
>>>
>>> My warmest regards,
>>> Maria Giatsoglou
>>> _______________________________________________
>>> Cmpl mailing list
>>> Cmpl at list.coin-or.org <mailto:Cmpl at list.coin-or.org>
>>> http://list.coin-or.org/mailman/listinfo/cmpl <http://list.coin-or.org/mailman/listinfo/cmpl>
>> <TestModel.cdat><TestModel.cmpl>
>
>
> _______________________________________________
> Cmpl mailing list
> Cmpl at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cmpl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cmpl/attachments/20160525/f2f9dc75/attachment.html>
More information about the Cmpl
mailing list