A functional framework for agent-based models of exchange
Résumé
We propose a functional framework for studying agent-based dynamical models of exchange. The framework consists of a notation and of a number of elementary definitions. We argue that - in comparison to narrative descriptions - the framework has a number of advantages: (1) it allows one to express precisely the relationships between the classical notion of general equilibrium and agent-based models of exchange; (2) it supports computational descriptions of models of exchange which can be unambiguously implemented; (3) it assists the numerical investigation of such models by providing a specific set of elementary notions and computational primitives. These can be used to deduce model properties and to setup "crucial" numerical experiments for validating model implementations. In a companion paper, we apply the framework to specify and study a model of exchange in which multiple equilibrium prices coexist and prices evolve according to a simple trading scheme and to a generic genetic rule.