Autonomy has switched over from subversion, over to git. The details for the new repository will be made available on the main page.

It is the year 4115, and the Earth government is under pressure; to maintain peace with the intergalactic neighbors over a deal that recently ended sour, they need to offer a half of what used to be known as the United States of America to a government team of investors from the nearby galaxy.

If they don't make a reasonable offer, specially after the recent blunder which forced them into this situation to begin with, a fierce war would be a real possibility - one that the Earthlings would want to avoid almost at all cost.

They don't want to solely decide which half of the the USA to sell, as they fear that if in time, these fierce investors feel that they've been taken for a ride (weather they have or haven't in reality is not even an issue), they may react in a damaging way.

They also don't want to offer more than a fair share for a couple of reasons - firstly, they're humans and naturally greedy, and would much rather minimize their losses, and secondly they don't want to appear weak and desperate (any more than they already have).

In other words, they need to slice the country in to two, fair slices.

Fairness in not purely in real-estate terms, but also factors in aspects of the land such as the weather, pollution levels, mineral resources, crime-rate, properties, etc.

As a side-note, it's also worth noting that what is worth one thing to the humans, is not necessarily worth the same to the investors, and vise versa.

What's the most elegant solution that the humans can offer, which will guarantee maximized satisfaction on both sides.

Note that while the aliens are fierce, they are still logical and would not wage a war if they can't logically prove that they've been wrongly done by.

Four arrow-headed agents are positioned to stand in a single line, all facing the same direction, but then the agent at the back of the line is then told to turn and face the other direction, taking the form illustrated below...

Next, 2 white hats, and 2 black hats are placed onto the heads of these agents, the order is completely stochastic.

Each agent cannot see anything behind them, nor can they see the the colour of their own hat. They can however see the colour of the hats of all agents that are standing in front of them (if any).

They are then informed that two of these hats are black, while the other two are white.

Their goal is to determine what colour hats they are wearing, and the only action that they are allowed to perform, is to state the colour of their own hat, once. They need to be absolutely certain of their decision as they are only granted one chance to speak. Each time any agent speaks, all other agents can hear what has been said, and if the voice is from in-front, or behind.

No other form of communication is allowed.

Given these circumstances, is is possible for the agents to correctly and logically deduct what the colour of their hats are regardless of how the hats are placed on their heads?

There is an island with 120 residents living on it, all friends, family... of which, one of them is crazy.

The island has 1 airport, and 1 plane, capable of seating 120 passengers. One day, the island leader decides to be extremely kind, and throws a party on a nearby island, all tickets purchased by him.

The people are obviously very excited, and come the day of the flight, people start to queue up to enter the plane. The craze girl, being the most excited, is the first in line, and the island leader having organized the event, ends up at the end of the queue.

Each person has their respective tickets, and each ticket has a unique seat number, however the crazy person couldn't care less which seat she gets. Each person that enters goes to their seat, in the event that their seat is occupied however, they then select the nearest seat to it and sit there.

Now, what is the probability that the island leader, being the last in the queue, actually gets to sit in his designated seat?

Games. Why are games so much fun? How is it that people can actually become addicted to games? Well, perhaps one answer is that they challenge our minds, to think, calculate, to be cleaver, sneaky and even a little deceitful. They don't need to be complicated, and even simple games such as Tic-Tac-Toe, Nim, or puzzles like the 8-Puzzle, 15-Puzzle, and N-Queens can be challenging at start. From there, games like sudoku, chess, backgammon, and go are difficult enough that it takes years to master. Nearly all games require some level of pattern-recognition, planning, compromising, judging of alternate moves, and of course, learning from experience.

Games are a playing field for development of A. I. ideas, and it is through games that many advances have been made.

One game that I think has much to offer for people interested in this area, but who also are in need of (but not interested in implementing) amazing graphics, is Unreal Tournament 3, Collector's edition.

If you're not a professional game company, but looking to get into game development, one way you can get experience is with the UT3 engine and tools. Just pick up a copy of the Windows version of Unreal Tournament 3 which includes the Unreal Engine 3 Tool Set. You could then build a mod for UT3, which can be freely distributed, to demonstrate your A. I. skills and ideas.

If you want to go professional you might be able to use that mod to attract a publisher to finance your engine license and distribute you game.

The Collector's Edition of Unreal Tournament 3 includes over 20 hours of video training modules to help get you started learning the engine. The non-collector's edition also includes the tools, but not the 20 hours of video instruction for them.

By the way - if you're going to make mods for UT3 you might want to consider entering them in the $1,000,000 Intel Make Something Unreal Contest which has recently been announced. See the  website for more information. The first phase of judging starts at the end of June but there are four phases over a two year period so there is plenty of time to learn the tools and get in on the fun!

You can also check out the Amazon exclusive UT3 developer interview with Jeff Morris on  Amazon.com.

Anybody interested in forming a group to work on this as a project, please register and express an interest by emailing me <nima at autonomy dot net dot au>.

Requests for, and Comments on Tutorials. Need help with a Topic? Blog it here!

One of the goals of Autonomy A. I. is to have a comprehensive collection of tutorials, along with any relevant source code. While Autonomy is still very young, and hence the collection is small, it is a perfect time to put in a request for a Tutorial.

If you find a tutorial that is incomplete, vague or difficult to understand, you're encouraged to make a note of it here because chances are that you're not the only one.

All tutorials are created right here.

The aim of this Wiki is to be well-referenced, accurate, and free of equivocal explanations and statements; anything that purports spurious information can be criticized. It's a very long way from being useful, but setting such policies now will help produce a more useful site in the future. The target audience for this site is at the moment, students studying AI, however as the site matures, this can and hopefully will expand to greater audiences.