Scouting smartly?

Started by
3 comments, last by AngleWyrm 10 years, 8 months ago

Update: thanks for all the no-reply, after some reading it seems that this is traveling salesman problem...

Advertisement

/Update:

wrote a simple recursive search, it's kinda fcking slow =d


        class rs_data
        {
            public List<ExplorerPoint> bestpath = null;
            public float bestpath_hdist = float.MaxValue;
        }

        private float recursive_search(float traveleddist, Vector3D lastpoint, List<ExplorerPoint> xplist, rs_data rs_data, int deep=0, List<ExplorerPoint> chain=null)
        {
            if (rs_data.bestpath_hdist < traveleddist)
                return traveleddist;
            if (chain == null)
                chain = new List<ExplorerPoint>();
            if (xplist.Count == 0)
            {
                if (traveleddist < rs_data.bestpath_hdist)
                {
                    rs_data.bestpath_hdist = traveleddist;
                    rs_data.bestpath = chain.ToList();
                    //backtrack last ?
                }
                return traveleddist;
            }
            var newlist = xplist.OrderBy(
                w => { return heuristicDistance((float)lastpoint.X, (float)lastpoint.Y, (float)w.Pos.X, (float)w.Pos.Y); });
            ExplorerPoint bestnode = null;
            float bestnode_dist = float.MaxValue;
            var i = 1;
            if (deep < 10)
                i = 2;

            foreach (var xp1 in newlist)
            {
                i--;
                if (i == -1)
                    break;
                var hdist = (float)Math.Sqrt(heuristicDistance((float)xp1.Pos.X, (float)xp1.Pos.Y, (float)lastpoint.X, (float)lastpoint.Y));
                if ((traveleddist + hdist) > rs_data.bestpath_hdist)
                    continue;

                var n = newlist.ToList();
                n.Remove(xp1);
                chain.Add(xp1);
                var res = recursive_search(traveleddist + hdist, xp1.Pos, n, rs_data, deep + 1, chain);
                chain.RemoveAt(chain.Count - 1);
                if (res < bestnode_dist)
                {
                    bestnode = xp1;
                    bestnode_dist = res;
                }
            }
            return bestnode_dist;
        }

Let me try to understand your problem, because it's not immediately clear from the pictures. You have a finite set of points (the red dots), you are currently at one of them and you want to design a path through all the other nodes, minimizing total distance travelled. Is that it? If so, ask Google about the "travelling salesman problem": The original problem includes returning to the starting point at the end of the path, but you can probably use pretty much the same algorithms in your case.

If that is not your problem, please try to state it as clearly as possible.

Never delete your original post and say "never mind, I've solved it". It just confuses people who come in later.

Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC
Professional consultant on game AI, mathematical modeling, simulation modeling
Co-founder and 10 year advisor of the GDC AI Summit
Author of the book, Behavioral Mathematics for Game AI
Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play

"Reducing the world to mathematical equations!"

Also, other people might have the same or a very similar problem. So solutions can be valuable to more than one person.

--"I'm not at home right now, but" = lights on, but no ones home

This topic is closed to new replies.

Advertisement