Course Approximation

I need a set of classes that can describe
1 a path as a
a) set of timestamps and acceleration vectors.
b) ellipse
c) set of waypoints
2 a waypoint as a
reference to a path and a specific time

The important thing i want to do is to cast a path as a c) to a a).
The reason i want to do this is to calculate the fastest path for a rocket which should be able to have waypoints as planets (b) and other rockets(a). The rockets themselves defy gravity and obey classical kinetics.
i know an analytical solution would be too complex so i can accept a reasonable approximation.

Anyone know where i can find something like this?

just dropped out of sight, up!

Quote:

---

Originally posted by kedaman
I need a set of classes that ....

---

What do you mean by that ? can you explain in detail ?

classes? The fundamentals of object orientation? Maybe you ask me to describe what kind of classes i need?
I'm actually looking only for this functionality, but as a sample set of classes:
abstract class path
class waypoint
class waypointset: public path
class vectorset: public path
Then when casting from waypointset to vectorset, the approximation is done, in another thread

LOL ! You mean the one We code in vb and save with a cls extension ?? :D

Actually I thought you were asking about Classes,instructors who
teach the subject !

HEHEHEHE :D
yeah actually c++

I am sure the guesses can be used to aproximate fast, but when it comes to several paths crossing each other i begin to feel inconfident. Depending from case to case i might get uncomfortable amounts of course alternatives due to permutations. Imagine Two rockets are travelling a set of courses, the first needs to intercept the second once, and both courses needs to be changed in favour for the general amount of time needed per rocket, if rocket 2 has to intercept other rockets on its way, the permutations problem becomes a fact.

The guesswork needs to be able to try a reasonable amount of alternatives, so that if some obvious solution would be present, it would probably also be choosen.

The user is not suppose to interact either.

May I suggest you consider looking at algorithms in FORTRAN (sorry). VB it ain't.

Specifically netbun.f and some others in fishpak
http://www.netlib.org/fishpack/

This is assuming I understand what you want: to figure elliptical orbit points that intersect given loci starting with an intial vector set (waypoints?). You will iterate thru large numbers of secomdary midpoint vector choices to optimize the path, ie., fastest arrival time with the slowest arrival velocity.... Elliptical PDE's I believe are worth considering. I haven't messed with this since 1996, so please take that as a caveat. Also, celestial objects move ala n-body problems. This problem is ignored for this exercise, I gather.

FORTRAN looks a bit too alien to me, thanks for the suggestion though

Quote:

---

FORTRAN looks a bit too alien to me, thanks for the suggestion though

---

Hehehe... :D

Kedaman: The general problem you have described looks overwhelming. I do not think you are going to be able to solve it.

Even if you work on a particular case with more than 4 intercepts, I do not think you are going to find a solution if you do not allow human interaction with computational programs. I am not sure you could solve a particular case with as many as four intercepts.

If I really had to solve this problem, I would first try solving a simple case requiring one intercept. Then see if I could use the technique in some fashion to solve a problem with two intercepts. Then try three. Then see if I could generalize the approach.

If I discovered that I could not find a general method applicable to any problem involving two intercepts, I would just give up.

Why not try solving a simple example of this type of problem? We should learn to walk before we try running.

I know, i made an algoritm for a single intercept with a lineary accelerating object, which was analytical, that was my first step, so the next is bigger than the highest mountain on mars. I wasn't even looking for trying to solve this myself, i just wondered if someone had done a similar aproach and if there was any libraries out there that i could have use of.
If not then i could either simplify this idea radically/ have the user interact by setting the course manually while the computer does the interception calculations, but that causes poor AI, or i could replace the whole idea and use either a warp or hyperspace model for travelling.