Thanks to scloopy I can make a start on this. I intend to feature the following as accessible objects in a compiled package. I want to build them so people can pretty much use them in a very abstract way. I lot of it is based on borrowed code so I will be rewriting a lot of it to make it more mine and more universal.

A*

Genetic Algorithm

Neural Nets

Especially with things like A* which has derivatives like Djikstra's, BFS and the sum of directions for the search there will be a lot of scope for variation. The neural nets I'm going to have to add one by one, probably starting with Back Propagation because it's the most useful.

Suggestions for other things I could add would be very helpful. I will add to this thread when I get the first basic components compiled and the documentation online.

Yeah, I'm running 5.0 anyway, but there are some issues with various things. Try searching the board for Java 1.42 or summink like that and there's various suggestions to certain bugs that can be solved by rolling back the Java. Not my problem so far.

Additions are more than welcome. I'm unemployed and trying to get some prints made for portfolio purposes as well as for possible exhibitions. I'm working on a piece right now to try to iron out the issues with the Genetic Algorithm library.

There's also stuff like the following links I got from talks in Prague I'd like to implement

Flocking behaviours and a **** load of AI links

Crowd motion

D* pathfinding (whatever the hell that may be - haven't got round to reading it yet)

Netlogo has a wealth of Artificial Life models on it. Some of the stuff like Ants I thought would have great potential for some art projects.

I am basically biting off my than I can chew here (gerbil trying to chow down a Buick if you need a scale comparision). The phrase, "anytime soon", really does not apply. But my major weakness is that I always make good on my drunken boasts. Eventually.

AStar Navigation Library

Pathfinder Library

This replaces the AStar library and is somewhat different to operate (and somewhat faster and easier). aStar(), bfs() and dijkstra() methods now exist to offer a range of navigation tools.

Nodes are now joined by Connector objects. This was necessary to make Dijkstra work because the Nodes in a Dijkstra search don't have a spatial location just a length of wire between them. However - the old Carden/Steed algorithm calculated distances every time it needed to find 'g', with Connectors it doesn't need to. That's a lot of calculations done away with.

I've also optimised the crap out of it. Nodes are located by simple x, y, z properties (no more arrays for location - they are slow). Pathfinder detects if Nodes both have z set to 0.0f and optimises distance calculations for the 2D user. ArrayLists - also faster. Home-grown indexing methods over Java methods - also faster.

I hope to build a kind of Bot class for it to make developing a way-point system for games easier. Moving less than a Node at a time requires some special trick. Right now I can't get my head round it and that means a week or so of developing time.

PS: Genetic Algorithm Library has had it's package name changed to ai.ga. The AI Libraries set will generally follow the ai.name system from now on.

I made a tiny variation of <Pathfinder_Dijkstra.pde> from the Pathfinder library.
It just put some colors on the connectors.

//from http://robotacid.com/
//from http://www.robotacid.com/PBeta/AILibrary/Pathfinder/index.html

// Demo of Dijkstra algorithm

import ai.pathfinder.*;

Pathfinder dijk;
int w = 10;
int h = 10;
boolean colored=true;
int spacing,s;
PFont font;

void setup(){
 size(400, 400);
 size(400, 400);
 spacing = width / w;
 s = spacing / 2;
 smooth();
 dijk = new Pathfinder(w, h, 1.0);
 font = loadFont("ArialMT-10.vlw");
 textFont(font, 10);
}

void draw(){
 background(230);
 fill(255);
 int m = (mouseX/spacing) + (mouseY/spacing) * w;
 rect((m % w) * spacing, (m / w) * spacing, spacing, spacing);
 drawConnectors();
 drawNodes();
}

void mousePressed(){
 int m = (mouseX/spacing) + (mouseY/spacing) * w;
 Node n = (Node)dijk.nodes.get(m);
 dijk.dijkstra(n);
}

void drawNodes(){
 for(int i = 0; i < dijk.nodes.size(); i++){
   Node n = (Node)dijk.nodes.get(i);
   Node p = null;
   int pnum = -1;
   if(n.parent != null){
     p = n.parent;
     pnum = dijk.indexOf(p);
   }
   int x = i % w;
   int y = i / w;
   int s = spacing / 2;
   if(p != null){
     int px = pnum % w;
     int py = pnum / w;
     int xDelta=px-x;
     int yDelta=py-y;
     strokeWeight(1.5);
     //choose connection color
     if (colored) {
       int rCol=0,gCol=0,bCol=0;
       if (xDelta<0) rCol=255;
       else if (xDelta>0) rCol=0;
       else rCol=127;
       if (yDelta<0) bCol=255;
       else if (yDelta>0) bCol=0;
       else bCol=127;
       stroke(rCol, gCol, bCol);
     }
     else stroke(10);
     line(s + x * spacing, s + y * spacing, s + lerp(x, px, .8) * spacing, s + lerp(y, py, .8) * spacing);
   }
   strokeWeight(1);
   stroke(10);
   fill(255);
   ellipse(s + x * spacing, s + y * spacing, 10, 10);
   fill(0);
   text(nf(n.g, 2, 1), 10 + x * spacing, 10 + y * spacing);
 }
}
void drawConnectors(){
 stroke(200, 200, 250);
 for(int i = 0; i < dijk.nodes.size(); i++){
   Node n = (Node)dijk.nodes.get(i);
   int cnum = -1;
   for(int j = 0; j < n.links.size(); j++){
     Connector c = (Connector)n.links.get(j);
     cnum = dijk.indexOf(c.n);
     int x = i % w;
     int y = i / w;
     int cx = cnum % w;
     int cy = cnum / w;
     line(s + x * spacing, s + y * spacing, s + cx * spacing, s + cy * spacing);
   }
 }
}