Modeling Organizational Adaptation as a Simulated Annealing Process

Abstract

Organizations can be characterized as complex systems composed of adaptive and intelligent agents. Organizational adaptation occurs through restructuring and learning. Organizations can be modeled using a dual-level model in which restructuring is modeled as a simulated annealing process and individual learning is modeled using a stochastic learning model and boundedly rational agents. Such a model is presented, and its behavior is illustrated using a virtual experiment where the type of organizational adaptation is varied. Results suggest that the organizational design and performance relationship may be chaotic, despite the simple rules of change. Simple restructuring rules lead to a wide range of emergent organizational structures that increases with individual adaptation. Organizations locate good designs (through chance and slow change) regardless of the agents' intelligence; however, emergent designs depend on adaptability. Design features are not systematically related to performance; rather, small initial differences in design and environment can affect the emergent behavior.