Multi-agent Systems

[deathtaker26]

Alright, this was an email I sent to my Programming professor. I'm going to post the email. Basically thumbs up if I got it, thumbs down if I don't. Please give me any information about it that you can. I am EXTREMELY hyped about educating myself about an M.A.S

> Professor Lara,
>
> I have taken the past few days researching Multi-agent Systems. I have to say I am a bit confused on the process. I wanted to make sure I had a full understanding of the system before I attempt to write my own. I'm not sure how familiar you would be with an M.A.S but perhaps you can help me out a little bit. I think I have a general understanding, but at the same time I feel that I am wrong about it. Can you tell me if my grasp of the concept is correct. If it is, can you maybe give me more information of benefits of using Multi-Agent Systems in programming?
>
> From my understanding, a Multi-agent System or an M.A.S, is a connection of multiple sources of intelligence acting and communicating simultaneously to solve a problem. The system basically tries to find the fastest, and most efficient algorithm to solve a problem with. When the system finds a faster step in its process or algorithm, it removes the slower step and replaces it with a new, faster, more efficient method. To summarize, it's teaching itself how to complete a task in the fastest time possible, almost as if a program is writing it's own flowchart to solving a problem.
>
> My example would be a program that solves a Rubix Cube. The first agent could take forever, but, upon completion, the program records the first algorithm it used to solve the Rubix Cube with. The next agent calculates another way to solve it. A step in the second agent's algorithm was faster then the first agent's, so it notifies the first agent that there's a faster step. The first agent replaces and shares what it knows with all the other agents. Soon, the agents collect enough data that it has the fastest method to solving the Rubix Cube. If all the agents come to the same conclusion as well, it is most likely verified to be the fastest algorithm to solving the Rubix Cube in the least amount of steps from the program's understanding with its given variables.
>
> If you don't know about Multi-agent Systems, that's fine, I will be able to read more into it. I just wanted a verification on my take on it.
>
> Thank you for any help in advanced,
>
> Dylan Dodds

–-Edit: Modified main post--