Computational Models of Human Organization Dynamics

The importance of modeling organizations and organization dynamics is becoming widely recognized. There is an emerging understanding that individuals do not usually solve large-scale problems, or necessarily even discover their existence (as individuals). Instead, organized groups of individuals embody the concepts, resources, and incentives that lead them to enunciate large-scale problems and opportunities and undertake appropriate large-scale responses. Our objective is to develop a computational model of organizations, suitable for capturing its objectives and structure, interpreting and predicting its behavior at various levels of abstraction, and detecting evolution of the organization's strategies relative to important objectives. Toward this end we have developed an organization model and an associated methodology that guides the acquisition of organization cases, interpretation of organizational data, and prediction of future behavior. We show that the methodology provides strong protection against modeling bias. All of this is illustrated in the context of a real example of organizational decision-making -- the decision by the political, scientific, and military elite to develop nuclear devices in South Africa. Anticipated Benefits: Any applications in which enough individuals are concerned with understanding and/or directing the activties of enough individuals, so that an explicit model of the organization of the latter individuals is required, will benefit. Commerical applications include businesses that need tools and methodologically sound consultancy services to get the right information to the right individuals in the organization at the right time (e.g., competitive edge, handle crisis situations).