Hypergame modeling of systems intelligent agents

摘要

We propose a new equilibrium concept in hypergames called systems intelligent equilibrium (Sleq). Hypergame theory is a game theoretical framework that deals with an agent who has each subjective internal view about a game situation and makes decisions based on it[3]. Sleq refines stable hyper Nash equilibrium[16] by considering some off-the-equilibrium plays. The main proposition suggests that it leads to Pareto-optimization from an objective viewpoint in hypergames satisfying some conditions. Practically and philosophically, the idea is inspired by the discourse of systems intelligence [7, 8]. We discuss the significance of systems intelligence in terms of hypergames and show some practical implications for agents acting inside a complex system. In particular we analyze paradoxical but ubiquitous social phenomena called systems of holding back.