Interaction dynamics for service design: simulating context sharing through collective improvisational dance

ABSTRACT: This study modelled service dynamics, specifically focusing on cognitive misalignments among actors, and conducted a multi-agent simulation using Bayesian inference, referring to an improvisational dance experiment. The results revealed that individual cognition influences context convergence: “No decay” condition fixed initial biases and hindered convergence, whereas faster decay increased fluctuation but enabled reconfiguration, suggesting the need for unlearning. When actors weighted others’ expressions less, cognitive misalignments widened despite strong subjective conviction.