Organizational adaptation to
ABSTRACT A computational model of organizational adaptation in which change occurs at both the strategic and the operational level is presented. In this model, simulated annealing is used to alter the organization's structure even as the agents within the organization learn. Using this model a virtual experiment is run to generate hypotheses which can be tested in multiple venues. The results suggest that, although it may not be possible for organizations of complex adaptive agents to locate the optimal form, they can improve their performance by altering their structure. Moreover, organizations that most successfully adapt over time come to be larger, less dense, with fewer isolated agents, and fewer overlooked decision factors. These results have implications for organizations of both humans and non-humans. For example, they suggest that organizational learning resides not just in the minds of the personnel within the organization, but in the connections among personnel, and among personnel and tasks. These results suggest that collections of non-humans may come to seem more intelligent (i.e., show improved performance) even if the agents remain unchanged if the system simply develops duplicate copies of some of the artificial agents and if the connections among agents are dynamically altered.
- SourceAvailable from: Omid Askari Sichani
Conference Paper: A Team-based Organizational Model for Adaptive Multi-agent Systems.ICAART 2011 - Proceedings of the 3rd International Conference on Agents and Artificial Intelligence, Volume 2 - Agents, Rome, Italy, January 28-30, 2011; 01/2011
- [Show abstract] [Hide abstract]
ABSTRACT: This paper proposes self-organization as a method to improve the efficiency and adaptability of bureaucracies and similar social systems. Bureaucracies are described as networks of agents, where the main design principle is to reduce local "friction" to increase local and global "satisfaction". Following this principle, solutions are proposed for improving communication within bureaucracies, sensing public satisfaction, dynamic modification of hierarchies, and contextualization of procedures. Each of these reduces friction between agents (internal or external), increasing the efficiency of bureaucracies. "Random agent networks" (RANs), novel computational models, are introduced to illustrate the benefits of self-organizing bureaucracies.
Conference Paper: Coevolutionary dynamics and agent-based models in organization science.[Show abstract] [Hide abstract]
ABSTRACT: This paper provides empirical and theoretical support for the application of coevolutionary dynamics and agent-based models in organization science. The support stems from the following logical progression: (a) organization science theorists have explored, and in many instances, acknowledged the applicability of complexity theory to organization science research; (b) much of the acceptance for complexity science applications follows from the conceptualization of an organization as a complex adaptive system (CAS); (c) complexity science offers a robust explanation of order in natural and social systems; (d) revolutionary dynamics provide the mechanisms with the highest explanatory power for describing order-creation in social systems. This paper provides an overview of the literature for each element of the preceding logical progression and concludes with a discussion of the applications of agent-based models to instantiate coevolutionary dynamics.Proceedings of the 37th Winter Simulation Conference, Orlando, FL, USA, December 4-7, 2005; 01/2005