Solver

[sawyerrnathaniel]

I kindly request your assistance with the attached file. I have bn trying to model the data using the solver function in excel but it has not made any success. I kindly request you have a look at it and see where I am going wrong.

[Richard Buttrey]

Would you explain all the terms and which cells contain the values. e.g. y(t), e,

[MrShorty]

Is your question about Solver or something else? Your solver model looks set up correctly, so that, with a good set of initial values, it should be able to find the "best fit" parameters. The problem seems to me to be in your modeling equation. A good logistic type function like this should have an asymptote (ym) as t goes to infinity, but your equation in column C seems to go on to infinity. I suspect that you need to spend some time with the formula in column C to make sure it is correct.

The first thing I might do is to pull up the formula evaluate tool (https://support.office.com/en-us/art...6-a70aa409b8a7 ) and step through one of the cells in column C. I might even have my hand calculator on hand and perform each step by hand to see if Excel is evaluating the formula exactly the same as I would. If there is any difference between what Excel does and what I would do on my hand calculator, explore that difference. With as many parentheses as you have in this formula, perhaps the first thing to be looking for in this step would be misplaced, missing, or extra parentheses.

[Alf]

Played around a bit with the Ym, U and the lamda values to see how the D51 value changed.

Based on this I set upper and lover values for Ym, U and lamda (range R11:S13) and used Evolutionary solver. Somebody with more mathematical
know-how than I should probably give you more sensible constraints / values than the one I have tested.

It also seems that the U value has very little influence on the result. I'v tested with min U = 0.00001 and U = 0.001 but the result for
D51 was the same in both cases.

It seems that putting the max values in range C4:C6 and then run solver gives a result a tad better that putting the min values in C4:C6
and then run solver. Why? I have not the foggiest.

Alf

[Alf]

Ooops! My bad, should be the other way round. Adding lowest values to C4:C6 gives a D51 of 34.93382

Starting with highest value in C4:C6 givesa D51 of 34.95674

Alf

[shg]

Assuming we're all looking at the Model 3 sheet, that's a very strange function to use to model a sigmoid. Where did it come from?