Use Case #1: Basic Scenarios of BDMToolbox

A typical scenario may look like:

Where active objects are:

1. Data Sources (bdm::DS), providing access to various form of recursive data
2. Bayesian estimators (bdm::BM), providing various exact and approximate algorithms for Bayesian inference
3. Controllers (bdm::Controller), generating control strategies, both feedback, or feedforward
4. Loggers (bdm::logger), for storing results of experiments

These objects operate on data vectors, functions and probability densities:

1. random variable (bdm::RV) is a name of a variable, used as named connectors in the scenarios,
2. probability density functions, (bdm::pdf), both conditional and unconditional,
3. functions, (bdm::fnc), of vector arguments,

These classes act as "LEGO-like" bricks which can be composed a wide range of arrangements.

Predefined scenarios implemented in mex are:

• Simulation, see simulator.cpp
• Estimation, see estimator.cpp
• Controll Loop, see controlloop.cpp
• Multiple-Participant Decision-Making, see arena.cpp

Lists of available objects are:

• List of Data Sources
• List of Conditional pdfs
• List of Non-conditional pdfs
• List of Loggers
• List of Bayesian Models
• List of Controllers
• List of Mergers

For details of their use, see tutorials:

• BDM Use - Probability density functions
• BDM Use - System, Data, Simulation
• BDM Use - Estimation and Bayes Rule
• BDM Use - Merging of probability distributions
• BDM Use - Control and Design of Decision Making Strategy