iteration

[SFares]

Hello,

Please scroll down in my sheet until you see the red arrow where i have my question. i am trying to maxamize the value for (Lpm_v) so that the ratio at the bottom does not exceed a certain value which i listed inside my sheet. Thanks!

[RichardJ]

Create a function, and then maximize it with a constraint using a solve block.

[MikeArmstrong]

I have plagiarized Richards method and shown a slightly different approach using a standard Given / Find solve block, as I am bored waiting for the WC.

In interesting to see that the two approaches converge to a slightly different result.

[Werner_E]

In which way is this a maximization?

[SFares]

Werner,

i am not sure if this maximization, but i want to calculate the maximum value for Lpm_v2 that will make the expression in the square root not to exceed the value that i listed.

[MikeArmstrong]

Sameer,

If you look at my worksheet I have calculated LPM_v2, as has Richard, but he has called in 'Answer'.

[MikeArmstrong]

Werner Exinger wrote:

In which way is this a maximization?

Who said anything about maximization?

[Werner_E]

Isn't it the value of L.PM_V2 which should be maximized according to sameer's description?

[RichardJ]

L.PM_V2 is something Mike added.

I just noticed that unless CTOL is reduced the result is not valid because it breaks the constraint slightly.

[MikeArmstrong]

It seems ok using Given / Find solve block.

[RichardJ]

I think Find always looks for "exact" answers, whereas maximize just gets to within CTOL, and stops.

[MikeArmstrong]

That makes sense

I have just increased CTOL and the differential between the answers reduces.

[RichardJ · Accepted solution]

I only just realized what you meant . I'm being dense today . Let's try that again (the answer happens to be the same).

[MikeArmstrong]

Not just you. I totally forgot about Maximize relying on CTOL.

[Werner_E]

As the function Comb_V_Pos_etc(L.PM_etc.) is strictly increasing, maximizing Comb or maximizing L.PM is the same.

Here is another slightly different approach using solve and the result is of course the same as Mike's. If you set CTOL to something low as 10^-7 in Richards sheet, his sheet also returns the same value.

[MikeArmstrong]

Werner Exinger wrote:

Here is another slightly different approach using solve

And the most elegant in my opinion.

Multiple variations for Sameer to work with.

[SFares]

Thank you all so much for your efforts!!

[MikeArmstrong]

No problem at all.

[Werner_E]

Here is another slightly different approach using solve

And the most elegant in my opinion.

I'll vote for Richard's second approach as sameer did. While more cumbersome its the only one which effectively maximizes L.PM and does not rely on Comb being strictly increasing.