Computational Modeling of Motivation
9.1. Notion of a computational model
Computational models have attained maturity in terms of perfection, variety and adaptability to numerous parametric situations. We find such models in various domains of computer science: artificial intelligence (AI – multi-agent approaches, artificial neural networks, probabilistic networks), theoretical computing (formal languages) and complex systems (self-organized models). Physicists have abstractly demonstrated a theorem of free will, and it is interesting to look at this theorem in order to understand the rational side of motivation. It is the only rational framework which deals with the issue. Computational approaches simulate a degree of reflexiveness, which is limited and is independent of natural language. All told, the formalisms proposed do not model self-motivation, but begin to approximate this concept by way of slight autonomy (internal motivation) or cost/benefit parameters (external motivation).
9.2. Multi-agent systems
One of the flagship formalisms to conceive of an artificial autonomy – which is, in itself, a model of a dynamic social network – is the multi-agent system. Ermolayev et al. [ERM 05], for instance, define actors and associated concepts in a multi-agent framework, which they call Dynamic Engineering Design Process (DEDP). Figure 9.1 presents the ontology of an actor in this model. An actor is the abstract representation of a person who performs tasks and executes ...