excel solver problem

interspaced · 15 Nov 2011 at 19:01

i'm trying to fit a 4-parameter curve using excel solver to this calibration graph only it doesn't seem to be working as intended.

Here is what it looks like before i run the data through solver

http://imageshack.us/photo/my-images/705/excel1.jpg/

And this is what it does with it

http://imageshack.us/photo/my-images/23/aftersolver.jpg/

Any ideas what i'm doing wrong

Makhaira · 15 Nov 2011 at 22:04

Not the faintest clue, but making the Excel file available to folks would probably be helpful

.

interspaced · 15 Nov 2011 at 22:29

i'm not so well versed in the free upload sites. What's good?

edit: found a site http://www.megaupload.com/?d=T85BHL2Z

Participant · 16 Nov 2011 at 14:00

Your 'solver' has just summed column D. What exactly are you trying to solve? Solver takes constraints and variables and works out the optimal result. I can't particularly see how that applies here as you haven't defined any constraints.

Interesting spreadsheet though! Chemistry degree?

interspaced · 16 Nov 2011 at 16:03

Essentially using that line on the graph and manipulate it so it fits the curve. The formula is a 4-parameter least squares regression so what i'm trying to do is to ask solver to find the minimum values for for sum of the d squares.

It's a weird method but it should work. It does work if you reverse the pH values then run solver.

edit: the constriants are the alpha beta gamma delta values that are supposed to alter the line when altered so it fits an accurate curve to the graph.