The Travelling Salesman and Chess

To do routing, timetables and scheduling, if you wonder how good artificial intelligence is give it the travelling salesman problem. Then we'll see how clever this AI is. Who says you cannot make money from mathematics never mind Graph Theory and Chess? Although I do admit at this stage my ideas are rather instinctive.

Computer Chess

The progress and increasing strength of computer chess was astounding, from the first ones that basically followed the rules almost just as random, until today where just an ordinary PC can play on grandmaster strength.

Obviously it is because of better and better machines but I believe the largest contribution is in software, as much more sophisticated algorithms and better understanding of the game and relevant analysing processes. This approach could well be applied in cyphers and codes too.

There are many possible applications, initially for very specific problems and beginning with simple, easily formulated ones like this travelling salesman problem. The problem is to find a journey that visits each city once, returns to the starting city, and minimizes the total distance travelled.

Logistics

Logistics and intelligence in peace and war is invaluable and essential. Teams of trained and experienced human minds can usually find satisfactory suggested solutions which I believe are usually better than existing software. Setting time tables and working out routes and in combination even just formulating a problem in itself could be complex. This is where “actual intelligence” comes in.

Often it is not necessarily critical to obtain the absolute best, the actual solution. Usually a good, a practical answer may be desirable and sufficient. You must have a solution if not then even a primitive one. Thing is you must have an answer, very good or not. We need practical and useful routes.

Of course there are known algorithms but they are not very sophisticated and are designed to find only the actual solution but not considering acceptable and useful ones. They are of little real use because of limited calculating speed and sheer enormity of the problem.

Applied Chess

Thus I suggest an approach adapting ideas learning from those in chess programs, crucial for analysing tactics and strategy from a given position including rules for evaluating positions. It involves complicated but very fast calculations analysing moves and possible positions ahead and choosing the best available move and then proceeding.

The current programs have taken decades to develop and by many different people it had a fascination from the start. The effort was incredibly successful.

Playing chess games and routing type calculations I believe work on similar principles imitating thinking as by humans. More of a hunch really, but a strong one, as finding good approximate solutions given the rules and other constraints, starting position and the actual stating of the problem.

I think they might work on similar ideas.

For our purposes you will in effect be letting the computer play against itself (or perhaps another?) Some reverse engineering would be in order too.

Finding answers

Your main applications would be of logistics and intelligence, I think mainly planning routes and schedules things that until now are mostly done by people. They have to be done and best we can is often enough so that I propose applying ideas of computer chess which by this time is incredibly sophisticated and efficient. 

The underlying ideas might be very successful in a much broader variety of situations.

In Practice

If applied as I predict and believe a terrific amount of money could be saved in general. I admit my ideas are still not clear and even vague but good software here potentially means megabucks.

The analysis and creating programs is too much work for one person, you would probably need a team of computer programmers and mathematicians, the necessary computer equipment and software to study and analyse the chess programs, and you would need the services of a patent lawyer and perhaps an engineer, and a grandmaster could help understand the ideas behind a chess playing program. And of course investment. Perhaps study other war games as well and why not?

I wouldn't try it on my own, anyway they will just pay me peanuts and think it's a big joke. At this stage it is little more than speculation, and I don't want to get caught up in something it sounds too much like work, but if you manage to do this and played it right you could get rich beyond your wildest dreams.