## 884 Reputation

16 years, 206 days

## Good....

@Carl Love thanks, I'll examine it tomorrow. It's bedtime in the antipodes.

## My Goal...

@Carl Love Thankyou, it works perfectly. How do you do it?!

This is to reflect that changing the order of the nodes changes the distance traversed. from my previous post:

tour_length([1,2,3,4]),tour_length([1,2,4,3]),tour_length([1,3,2,4])=85,88,73

inspired by:

where I was unable to write the ILP constraints properly. Hence my need for the sledgehammer exhaustive approach above. With the new distance matrix:

VRP_IP_DC1.mw

You can be the first to answer this question!

## Inds...

@Carl Love thanks for your effort.

Sorry to bug you further. How do I return the row numbers of the matrix <LimitDist(56)[]>;

[[1, 4], [1, 5], [1, 2, 3]] = inds 4

## You interpreted correctly...

@tomleslieJust what I wanted! I salute you Sir and Carl Love too.

BTW this problem is inspired by an earlier unanswered question,

where I was unable to write the ILP constraints properly. Hence my need for the sledgehammer approach above.

## Yes, but...

@rlopez the OP asked specifically about output of Std. Error w.r.t. NonlinearFit.

"the summarize option is not available when the model expression is not linear in the parameters"

## It doesn't...

only

The summarize option returns a summary for the regression:

Edit.

## Sen?...

@SamuelPita well where is the text or .mw file? Solving 5 equations for 5 unknowns should work...

Try substituting sin for "sen".

## Robert,...

I saw this problem a while back on http://www.qbyte.org/puzzles/p004s.html but I couldn't understand their explanation.

using the big green arrow. no-one here is masochistic enough to transribe code from a picture.

## Outstanding...

@vv how did you derive the expressions for R and d?

## Thanks for that...

@Kitonum . Code with annotations too, bonus!

## lateral thinking...

@tomleslie thanks. And vv.

## Numeric solution only I think....

You're out of luck. k=x+e*sin(k) doesn't have a closed form solution in terms of k. If you supply values for e and x, you can find k numerically using fsolve.

@Kitonum and Tom