Language selection
 English language   French language 

 
You are visitor
 0 2 4 9 4 3
 
Search
 
Theme selection

 
Last Blog subject
Analyzing ...
Parameters Tuning Unity
Domain
Application (artificial intelligence, real-time, client/server)
Context
Realization of an application in Open source during my spare-time
Period
Mayo 2004   -   January 2005
State
Finished (and maintained)
Language(s)
Java
Link(s)
Development's site  -  source code

Description
I have produced the first version of this application, named 'Parameters Tuning Unity' (PTU), during my research works about the Parameters adaptation into reactive multi-agents systems: application in mobile robotic, whose it can be interesting to consult information.

   The application's aim is to give a support in the resolution of the issue which consists with the adaptation of the parameters of an expert system or a multi-agents system (which might be expendable to other types of systems).
For that, it lays out:
   - a client/server architecture in order to communicate with the system whose we want to adapt the parameters,
   - an extensible 'range' of optimization's methods (such as the gradient's descent ).

Basically, the principle is relatively simple and can be regarded as the following use's protocol:
   1 - creation of an XML file defining the characteristics of the parameters to adapt, and a stop's criterion,
   2 - initialization of the PTU;
         - call i the index of the 'current' parameter, initialized with 0,
         - initialization of the parameters with their initial value.
   3 - initialization of the system,

   4 - the PTU sends the current value of the parameters, to the system,
   5 - the system "lives a cycle" then sends an evaluation of its state (which is the one to optimize) to the PTU,
   6 - the PTU uses the optimization's method selected by the user, then updates the value of the parameter i,
   7 - update of i itself,
   8 - the PTU checks the stop's criterion,
         - if it is not reached, return to 4,
         - else end of the optimization's process.

   9 - analysis of the results.

The optimization's methods are implemented as modules, and can thus be easily added (or removed) to the application. The PTU's evolution so consists mainly on the addition of such modules. Its architecture is currently under update, and a simple example, simulating an expert system, is under development, in order to highlight mechanics presented above.

Remark(s)    
Whoever are you, if you wish to take part in the development of the application, do not hesitate: mail me.

Installation
Consult the 'README' file provided with the source code.

Copyright ©2003-2009 Bertrand BENOIT,  Linked LinkedIn profile profile.
Created the Sunday 05th October 2003. Last update the Saturday 21st November 2009.

You can download my Curriculum vitae.