0

Excel's Solver is telling me "the linearity conditions required by this LP Solver are not satisfied." It's lying! Does anyone know what the linearity conditions LP Solver requires actually are?

I suspect this is a formatting issue. I've searched other Q&As, but most issues other people seem to be having are case-specific, as is mine. I've not found the actual conditions clearly stated.

https://drive.google.com/file/d/0BwIofgnF8b9Ndlczb0tMQzN3QUU/view

Above is a link to my document. From there, if you download the file and enable editing, my objective and constraints should still be saved in Solver for your ponderment. In case they're not:

Set Objective - $I$387
To - Max
By Changing Variable Cells - $F$2:$F$201
Subject to the Constraints - $F$2:$F$201 = binary
- $H$387 <= 493.45
- $J$387 = 2
- $K$387 >= 4
- $L$387 >= 5
- $M$387 <= 2
- $N$387 <= 2
- $O$387 <= 2
- $P$387 <= 18

The sheet I'm concerned with is the last one. I've included the linearity report in the sheet prior. The report indicates that nothing is linear. So, I suspect there is a fundamental error in the document, perhaps in the way it's formatted.

Thanks for any and all help!

DataProphets
  • 156
  • 3
  • 17
  • Why not use GRG? How do you think formatting will lead to nonlinearity? It's not likely someone will open your google drive doc. No permission, risk. – OldUgly Feb 19 '17 at 03:03
  • GRG "cannot find a feasible solution." However, manually, I was able to find a perfectly feasible solution myself. This is why I'm confident the error message is not, theoretical, accurate. It must be a more trivial error, such as improper formatting, a typo, or, perhaps, a factoid about the properties of the Solver I'm not aware of. Is there a better way to make my issue reproducible? – DataProphets Feb 19 '17 at 06:16
  • This can happen if the model is non-linear and non-convex. If your model is linear the Simplex method is more appropriate. To my knowledge Solver is usually correct when it says a model is nonlinear.. – Erwin Kalvelagen Feb 21 '17 at 20:36
  • I resolved the issue. Some of the cells contained a formula referencing other cells whose formula contained an IF function. This IF was set to return 0 in one case and the text from a third group of cells in the other. By deleting the middle group of cells and directly referencing the third from the first, I was able to use IF functions that returned only numerical values, and never text. This seemed to make the difference. So, while the model was linear from a theoretical perspective, my lack of intimate procedural knowledge meant that a formatting choice made it not so in practice. – DataProphets Feb 22 '17 at 00:33

0 Answers0