An Agent-Based Model of Miscommunication in Complex System Engineering Organizations

Abstract

Communication in engineering organizations affects the performance of the complex systems they design. Miscommunication occurs when communication results in a “deficiency” or “problem” that hinders parties from fulfilling their individual or collective values. A recent study demonstrated widespread miscommunication in a Fortune 500 engineering firm about the definition of “an estimate” in a complex system design context. Building on that work, this article used a Monte Carlo simulation (8800 runs) of an agent-based model to demonstrate how systemic design process miscommunication may affect complex system performance. Each run of the simulation created a unique 1000-artifact system using a network generation algorithm and converged its design through optimization. Systems where estimates communicated “current” designs outperformed systems where estimates communicated “future” projections of their designs instead. Varying the fraction of the population, which uses each definition of an estimate varied system performance and uncertainty. Whether related to estimate definitions or more generally, this article demonstrates that miscommunication may affect system performance.