Abstract: The nature of the organizational boundary is investigated in the context of organizational learning. Boundary permeability is defined and hypotheses relating it to performance are tested computationally using data from 5,500 artificial organizations. We find that matching boundary permeability to the environment predicts both agent and organization survival.
Abstract: A theoretical representation of social structure in agent-based organizations is developed. To test the model we generated a hypothesis from organizational learning theory and tested it using computational experiments. We found that emergent social structure associated with rewarding agent learning increased collective output over and above pay for performance.
Abstract: Using the computational technique of agent-based modeling, we examined the ways boundary-spanning agents influence an organization’s learning capability under various environmental conditions. We propose a model wherein an organization is represented as a network of connected agents, tasks, resources and knowledge (Krackhardt & Carley 1998). Boundary spanning agents searched for and retrieved new information from across the organizational boundary (Richardson & Lissack, 2001) and returned to apply the information as task knowledge and to diffuse it throughout the organization. The Organizational Learning Systems Model (OLSM)(Schwandt 1997) was the basis for measuring outcomes at the organization level. We found that:(a) the number of boundary spanning agents was a predictor of collective survival in changing environments and (b) production was higher and fewer boundary spanners were necessary for higher levels of output when agents learned new tasks. We conclude with a discussion of the broad potential for this model and these computational techniques.