Bounded Rationality

Abstract

Abstract This chapter follows the introductory chapter and serves as a fundamental step in agent-based modelling (ABM), focusing on the concept of bounded rationality (BR). As a nonstandard review of this otherwise hoary subject, this chapter argues that bounded rationality is a natural and inherent aspect of ABM, making ABM a sufficient framework to justify the concept. To do so, we distinguish ABM from the conventional equation-based modelling (EBM) and use the mirroring principle to argue that it is easier to have perfect rationality in a bounded model like EBM but not an unbounded model like ABM. The connection between ABM and bounded rationality is illustrated using two canonical kinds of ABM, namely, evolutionary computation and cellular automata. Apart from the mirroring principle, through the introduction of several other key principles, including the modularity principle and the entropy maximization principle, we advocate the idea that bounded rationality, while motivated from multiple perspectives such as behavioural economics, biology, psychology, artificial intelligence, cognitive science, and physics, can be further advanced through ABM. ABM provides a cohesive framework for interdisciplinary discourse among these fields.