I'm outa here... vacation, then a conference

[Mark_Everson]

I will be gone, with only intermittent net access (and no email) for the following dates:

July 1-10, and 14-17.

I'll try to check on the forums when I can.

We seem to be falling a bit off our planned schedule . I am one of the transgressors. I had an outbreak of work-related activity and a bit of 'designer's block' on the Econ system.

However, I'm really feeling like getting an alpha version out asap. Also, I think that although we are lacking in many details, things as far as they're sketched out seem to fit together reasonably well. So I propose we stick to the original process and dates, and just have things in some areas a little less fleshed out if necessary.

In my absence the dukes IMO should also think about a first, relatively simple sub-set of their areas that we can implemented for the first alpha. Once we have those specs and the variable / interface information that goes along with them we can start coding on alpha1.

I'll finish the econ model on vacation, and try and take a first stab at a high-level coding spec.

Have Fun,

-Mark

[JimC]

Progess report:

I have a Genetic Programming system nearly completely up and running and very thoroughly tested.

Once we have a game engine working, I can start AI coding in earnest.

I'll try and get a Java applet up and running on my webpage ASAP demonstrating a simple GP problem.

Reminder:

My Page

Map AI Applets are also located here.

Jim

------------------


[This message has been edited by JimC (edited June 30, 1999).]

[MarkG]

testing something....

[Mark_Everson]

Jim:

Brainstorming on Map AI. What do you think of the idea of using a graph (in the 'connected nodes' sense) to streamline movement AI? So for instance if Africa were a single region it would be a node with connections to the Middle east thru the Sinai (neck / hotspot) to the Mediterranean thru the north coast, to the Atlantic... There would need to be some information about the size of regions and connections (necks)

So when going from Cape Town to Rome the movement AI's first step would be 'What are the possible paths?' Those might be in words, go to Atlantic, Atlantic to Mediterranean, Med. to Rome. Another would be overland to the north coast of Africa, then to Med. to Rome. This kind of analysis would reveal immediately that some paths are impossible due to firm enemy control. The A* search could then be pruned by eliminating areas that the graph says are impossible. Alternatively a GP could be used to evolve some rules of thumb for expediting path-finding using the graph information.

I'm not sure about the feasibility of this approach, but it seemed worth bringing up.

-Mark

[Mark_Everson]

Well, I'm off to my AI conference. I'll be back checking in on 7/20.

-Mark