Why the decision tree approach? (14-05-03)

1. Discrete approach

The decision tree is a discrete process. This is the main reason why it has been decided to use such an approach. Follows a definition of a discrete process (from continueProcess): By discrete process, according to the input point of view, it is meant that inputs are interpreted in a discrete way. To understand what they mean, conditions are checked. Fixed values are used. A condition is satisfied or it is not. No intermediate value is accepted. According to the output point of view, the output values are discrete and the set of outputs is limited.

All the information gathered, all the knowledge acquired provides discrete values. Consequently, this set of discrete values has been completed (as it has been written many times, a lot of information is missing) so that every set of parameters value can be compared to this one. Then a situation is recognized and a decision is taken.

Thus, considering that, a discrete approach is more convenient.

2. Simple approach

Real time constraint: since the bot evolves in a real time environment, the making-decision process has to be quick and light enough to respect this constraint. Since decision trees provide simple and quick algorithm, they are well designed for real time reasoning.

Simple and well-known structure: the tree structure is very convenient to organize data, and especially in that case since it is leveled structured: each level in the tree is related to one parameter. Moreover the tree structure looks familiar, thus it is easier to build and use it.

First approach: since this work stands as the first step of a flight simulator bot design, the decision tree approach is well appropriated to be use as a first one. Up to the following researchers to improve it.