Week 1 Complete

Well after the first week only Martin and myself have submitted our AI. We both solve the 3 demo levels correctly. Martin seems to be ahead on the path optimisation as he seems to have the shortest path worked out to avoid making unneeded direction changes. Mine however is just a simple breadth first search which doesn't take into account the turns. In fact it's still in hack stage because it only checks to see if the box can be moved to the goal it never checks to make sure the player can push the box to the goal as well.

Excuses: Andy has told me he's starting this weekend. Gordon has got it compiling now and is investigating a C version. Mayo promised me a demo AI, where is it???? Ben has started creating his AI on paper, I'm not sure if I'm going to get a paper version sent to me yet. Martin has some PhD friend who has also entered CodeName "GebAI" apparently his AI also solves the 3 demo levels in an optimised way. Simeon has converted the code to C# and is investigating that. The others out there are either too scared, or are secretly developing their ultimate AI.

I've now included the 90 semi official levels for you to test with. My AI cannot solve any of them yet. The best research AI out there can solve about 2/3 of them. Let me know when you can solve one of them and I'll start up a page for them too.

The demos for week 2 are a bit more complicated, the order of the moves is more important now as is choosing the correct goal and also the correct path to get to the correct goal.

Week 2 Complete

Well actually this week has only begun. No, wait now it's ended.

This week Gordon has made a stunning entry into the race, he's taken the lead after much cursing at Java. I'm very interested to see his code at the end because his design seems quite different from mine. Gordon was the only one to solve all the demo levels. His code does get a memory exception when trying the first official level so I'm happy with that as it means he has some more work to do.

Excuses:Martin has been pretty quiet the past week. I guess he's writing some patient for the secret service that he works for. Andy seems to be pretty quiet too, I still haven't seen anything yet Andy. Mark was full of hot air once again. Where is it Mark? Everyone else is also keeping quiet on their designs. I've been doing a bit of work on mine but was away for 4 days running in Scotland. Also I have a couple of bugs which are stopping from solving the last two demo levels.

Next week I think we'll include the first official level to see if anyone has progressed that far yet.

Week 3 Complete

Not much activity this week, I've been busy with work and other things. This weekend should prove a little more productive since I think my finger is broken so I can't go climbing.

Martin has informed me that he has big plans for next week. Gordon seems to be busy improving his code, I've allowed the use of the following 3rd party code at the request of Gordon. Minimum Bipartate Matching (Hungarian Algorithm)

The levels for next week have been provided by Gordon.

Week 4 Begins

Well actually this week has only begun.

Files

Links