A pair of pictures to start.



By now I'm sure you're seeing the relation and trying to see whats so special about it. If you've figured it out by now, I bow before you. However, for those of us who aren't so damm freaking smart, a helping hand for you.




And by now I'm sure you're all thinking I've lost my mind, but fear not! For I have created a Dijkstras pathfinding algorithm using basic COG for JK!! I have included a zip that contains a gob of my little test level which I run the little uggie through, and you will note that it is completely dynamic. (You'll find his start point through the 1st corridor you see when you boot it up) However, I must admit, it still has some flaws, one of the nodes has some double return stuff which is pretty trippy....cycadelic 70's doublevision maybe lol. Its currently designed to run the guy around to random target nodes, but I dunno whats going on bug-wise yet. However, the concept is valid which is the showcase here. Me shutup now. Any questions/flames/curses of eternal damnation for making you want uggie-babies, post them.

And I still own copyrights to this phrase....
So whatcha think?

[edit]Oh and those other two AI's there are merely there b/c my laziness to remove them and my other little AI cog coding that I'm currently tweaking with...[/edit]

A diji-what now?

I think he made an ugnaut that goes along a preset path, but it stops when you get too far away so you can catch up.

Awesome, nice work.

For those who are confused:
http://en.wikipedia.org/wiki/Djikstra%27s_algorithm

Nice.

The old girl might have some life in her yet.

Every computer science and software engineering student at RIT implements Dijkstra's algorithm within their first two quarters. Don't feel so smart.

So?

Heaven forbid he enjoys doing something new in the context of JK's COG.

Emon - WRONG. You do it in CS3, which is traditionally third quarter.

Also: that's pretty groovy.

You used to do it in CS2 before they switched the curriculum.

Echoman, it's not about JK so much as d_p seeming to act smart over what is a trivial problem.

But ignore me, I've had a bad day and am making comments without thinking.

Could come in useful. Don't make me work it into another SP level...

Okay, some updates on the coding, I've manged to get it to run through up to 3 cycles at times, but most times only 1, what happens is it finds the path and then just does nothing, it perplexes me exactly why this occurs, cause I don't think its too many threads running the cogs, but *sigh* idk why it does this sometimes for some things and not for others. Also, with my cogs for the nodes, when I reorder the children, node 1 operates just fine, but otherwise still with the double calling stuff, and I think a few other nodes do this too since they have weird children, but I haven't tested explicitly yet. However I'm posting the updated gob and code and it will pick a node at random to start off at, just to prove that its dynamic.

One thing to note, rand() gives me probs such that everytime I restart JK completely, it seeds the rand() the same way, so what you need to do is restart the level in-game and it will travel to another node. Feel free to open the gob and look at my code and level and find whats going wrong with the execution (cause . *grumbles and sighs*

And on a cogging note, apparently when you call SendTrigger() it pauses the execution in the code (up to a sleep()) probably so you can use a ReturnEx() like for SendMessage[Ex](); and call foo;

Okay, perhaps abit long-winded.....However, Emon, you do it then. Yes the algorithm is simple, but doing it in C++ or other true prog-lang is sh!!-mofo easier than in COG.

Get Zeq and Co. to hack in a call to seed the random number generator.

It's just A* without a heuristic. Why do people make such a big deal out of dijkstra?

Cause Dijkstras guarentees the best solution by searching all possible pathways and is one of the fastest known algorithms to do so. A* just gives a very good looking solution, which might (tho unlikely) be the best. And I've heard Dijkstras is alot easier to code than A*, but I dunno on that one *shrugs*

Why are you talking about "child nodes" and "reorganising"? You should be using a graph data structure, not a tree.

Also, what's with this "double return stuff" that one of the nodes has? I don't quite understand what you're talking about there.

Well, how I have it set up, I have a much of cogs which hold the nodes data and I send a trigger out to retrieve that data from those cogs. Each node comes with 4 "children" nodes to account for 1 way pathways and such, so its not really children per say but the edges to adjacent nodes, oh and they are numbered 0-3, so I mean I take child1 and swap that with child3, ie permutations of the children order.