hello everyone!

im a computer science student, i have studied multiagent system and i have the idea of applying multiagent idea on operating system so it's behavior can be optimized via making each unit in OS as an agent collaborated with other agents to take there decisions according to there and others knowledge .

and to take there decisions according the situation, and history.

 

i want to apply this concept on part of linux , like on memory management part.

 

my supervisor likes the idea.but i wonna know for how much i can apply it and what are the good tools can be used for? im studying multiagent systems now, what references i can use ? for details please check the report

http://www.4shared.com/file/142130574/125092c0/Smart_operating_system.html

I'm not entirely sure what you're proposing. Are you suggesting a system somewhat like GNU Hurd, but with more "intelligence" in the servers?

 

(Hurd's concept is to split the operating system into various "servers", which operate independently.)

In what aspect will it "optimize OS behavior"? Could you give us examples.

---

Merged post follows:

Consecutive posts merged

I just read your report.

The things you are describing as goals for the multi-agent "context sensitive" OS, are already implemented, and by much simpler means.

Quick examples:

- fingerprint recognition is implemented in many laptops, intended to protect sensitive business information.

- Windows sets system recovery points before new driver installations, and i am sure SuSe had something similar too.

- speech recognition is done very well on Windows 7 - the user is being able to command almost any software, using the speech recognition engine.

 

There are some other generally good ideas in your report, but I think each of them could be implemented by individual software, rather than everything done by one multi-agent system.

 

Now, memory optimization sounds interesting, but how can you apply the idea of MAS to it?