Rody Oldenhuis

Oh no!!!! :)

Anyway, I suspect it's the global optimizer: Although the solution I reported gave my ~15km/s, that does not mean that solution is easy to find! In my thesis there's also a section of "raw data" (somewhere in the appendices), which shows per point I tried that different optimizers often find wildly different results.

I suspect that if you retry a few times, it'll probably decrease from a few hundred to a few tens to (hopefully for me!) ~fifteen.

Anyway, keep in mind that the global optimization algorithms I used (GA, PSO, etc.) use a fair bit of randomness. Chances are you'll find different results each time you run an optimization.

Of course, if you try 100 times and you still don't find anything lower than say 25km/s, I'd like to hear about that :)

I can think of at least one really useful application for this: to write sloppy, quick-n-dirty, messy, one-time use only, it just works®, thinking-while-writing, deadline-is-near, just finding something out kind of programming :) I know most people here do that a lot, admit it! So instead of indirectly denying this and ban all functions that help and facilitate this kind of sloppy programming is a bit short-sighted in my opinion.

Of course you're not going to use it in any of your "real" work, stuff that others have to see: 1) the author has given a warning as to why you should not do that, and 2) if you still do it, you'll find out soon enough why GOTO() isn't so great.

To give this submission one star because you think people will use GOTO() in "serious" functions or "teach bad habits" is just not recognizing what this function was really made for (or will likely be used for). To jump straight to your purist programmer's saddle and scream fire and death to GOTO() is just silly.

It's actually a pretty good job for what it is. I give it 4 stars.

Done!

nice, but change line 12 in the testoptimize
to :
clc, rosen = @(x) (1-x(1))^2 + 105*(x(2)-x(1)^2)^2 -6000*sinc(x(2)^2+x(1)^2); % adding local minima.
and you get a local minimume solver as fminsearch is.

Hi,
I generated two plane with as follows;

par1.a=2;par1.b=2;par1.c=1;par1.d=2;
par2.a=10;par2.b=1;par2.c=2;par2.d=100;

P1=generate_plane(par1,-10:10,-10:10,'r.');
P2=generate_plane(par2,-50:-40,-50:-40,'b.');
---------------------------------------
function P=generate_plane(par,rangex,rangey,color)
[x y]=meshgrid(rangex,rangey);
z=(par.a*x(:)+par.b*y(:)+par.d)/-par.c;
z=3*rand+z;
P=[x(:) y(:) z];
----------------------------------------

Then find the principal axes and orientations as follows,
[U1 E1]=svd(cov(P1));
[U2 E2]=svd(cov(P2));

a=[E1(1,1) E2(1,1)]; %principal axis
b=[E1(2,2) E2(2,2)] ; %minor axis
c={mean(P1), mean(P2)};
u= {U1(:,1)', U2(:,1)'};
v= {U1(:,2)', U2(:,2)'};

However the algorithm can not find the correct distance. Also if two ellipses are intersect it fails. The result should be 0.

For 3D, an explicit example is necessary.