As suggested by the name, Mixtools 3000 is designed as a successor of the successful package mixtools. Original mixtools were designed for the purpose of Dynamic Bayesian Advising, however their structure does not allow for multiple participant decision making which is a new area of research.
The main aim of the project is:
In order to achieve this goal, the software must be flexible, easily extensible, and user friendly. This can be achieved via careful analysis and implemenation via object oriented approach. Local heritage of Matlab code and Matlab-centred working enviroment forces us to use Matlab as a user enviroment. On the contrary, we need fast and numerically stable code and implementation of basic features of object-oriented approach. This is achieved using our home-bred set of macros which allows seemless integration of Matlab, ANSI C and elementary support of object oriented approach. For more information see our technical report.