Hi,
I know this is very late, but I just want contribute for anyone else with the CPLEX Error 1233... since there is still little documentation. Once I received this error, I wrote a code that checked all parameters I have as for cplex, to see if they are
within the numerical bounds. (FYI I thought the problem was with my sparse A matrix)
Problem and solution: I had lhs and rhs vectors containing -inf and +inf, and as it turns out I accidentally placed a negative in front of a set of my rhs inf values. Cplex told me that the problem was at index 108, however the index of this error was at
109 of my rhs. Once I fixed this, everything was fine.
SO to those who encounter this error and don't know what to do, try writing a small function what go through all values for your input parameters (e.g. rhs, lhs, varUB, varLB, and A) that will stop once the problem is found. Also don't forget to check
both the index and parameter values if you are using a SPARSE A matrix.
Best,
Cynthia
On Wednesday, June 13, 2007 at 2:34:11 AM UTC-4, csviks wrote:
Hi:
I am a newbie to CPLEX and was wondering if an expert here might give
me some advice on how to debug this CPLEX error. I am trying to solve
a linear program of size (cols,rows) = 1298432 3358720, the size is
large but it is quite sparse with each row having no more than 1000
non zero entries. My first question is if I can expect CPLEX to solve
it. Currently, I am unable to get to the stage where CPLEX can solve
it because of the following error.
CPLEX Error 1233: Numeric entry 903180 is larger than allowed maximum
of 1.79769e+308.
I know for sure that the non-zero coefficients are definitely less
than 5 though the indices of the row and columns could be as high as
above = (cols,rows) = 1298432 3358720. I tried looking at the
reference manuals but the explanation of this message is quite limited
as follows.
CPXERR_DBL_MAX
1233 Numeric entry %d is larger than allowed maximum of %g.
Data checking detected a number too large.
Can somebody suggest how to go about debugging this? What does numeric
entry mean? Is it the "value" (coefficient) or simply the "index" of
the row or column? But then, the allowed maximum is far more than this
size.
Any help/hints will be appreciated.
thanks.
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)