(Emotional) Learning Classifier Systems ((E)LCS)

Learning Classifier Systems (LCSs) are a kind of a specialization of the general concept of a GA because within LCSs the population of 'genes' is formatted as a population of classifiers. A classifier (cf. figure ) is a unit combining two parts: an IF-part -also called condition- and an ACT-part representing some kind of an action'. The whole population (P) of classifiers is working in a classifier system as a mapping of the kind

$$P_{CLASSIF} : INPUT \times I_{SYS} \longmapsto OUTPUT$$ (4.1)

where a set of possible input-vectors is matched with the IF-parts of the classifiers. Those classifiers which match are then candidates for the generation of an action. The final selection of an action depends from some additional internal states $I_{SYS}$ of the system; usually are this the actual strength-values of the different classifiers.

From another point of view are LCS like agents with input and output whose internal behavior-function $f_{LCS}$ is based on a set of classifiers which are managed by an evaluation function $eval_{LCS}$ which applies a genetic algorithm onto the set of classifiers by exploiting some feedback (or fitness) values depending on the success of the behavior of the agent in some environment.

In the classical versions of LCS systems (see 'Some Bits of History' below) the fitness values have been deduced from the environment without including the learning agent. Therefore was the LCS theory somehow unfinished or incomplete. A major factor of the learning process was somehow 'outside' of the scope of the theory. Everybod who did use an LCS system had to provide it's 'own' model of fitness generation.

In recent years one can observe a shift in the way how to look to the fitness values. In Neuropsychology e.g. (c.f. Solms et al.(2002)[] one could work out that the body itself has several basic processes managing survival which communicate with the brain and which are signaling built-in fitness values for a basic orientation. And this new concept of 'built-in fitness values' has already found its way into robotics (see e.g. the paper Gordon et al. (2010)[111]). But there were several precursors to Gordon and the label of emotional robotics for this is already established too (cf .....to be done....).

Combining the classical LCS paradigme with this new approach of using built-in fitness values we will construct the LCS systems (starting from level $ANIMAT^{2}$) with such kinds of fitness values and call them therefore Emotional LCS (ELCS).