A Physics of Organizational Uncertainty: Perturbations, Measurement and Computational Agents

Of two contradictory approaches to organizations, one builds a plausible story to advance knowledge, and the second resolves the interaction into a physics of uncertainty to make predictions. Both approaches lend themselves to computational agent models, but the former is based on information derived from methodological individualism operating within a stable reality, the latter on interaction uncertainty within a bistable reality. After case studies of decision making for Department of Energy (DOE) Citizen Advisory Boards, we relate the lessons learned to an agent model (EMCAS). We conclude that simple interactions can produce complex stories of organizations, but poorly anchored to reality and of little practical value for decision making. In contrast, with a physics of uncertainty we raise fundamental questions about the value of consensus to instrumental action. We find that by not considering uncertainty in the interaction, the former model instantiates traditional beliefs and cultural values, the latter instrumental action. IDEA GROUP PUBLISHING This chapter appears in the book, Computational Economics: A Perspective from Computational Intelligence edited by Shu-Heng Chen, Lakhmi Jain & Chung-Ching Tai © 2006, Idea Group Inc. 701 E. Chocolate Avenue, Suite 200, Hershey PA 17033-1240, USA Tel: 717/533-8845; Fax 717/533-8661; URL-http://www.idea-group.com ITB11872 A Physics of Organizational Uncertainty 269 Copyright © 2006, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. INTRODUCTION: ORGANIZATIONAL SCIENCE According to organizational theorists (Werck & Quinn, 1999), the state of theory is poor and little advanced from the time of Lewin (1951). There are at least two broad but incommensurable theories attempting to move theory forward: Traditional methodological individualism, which tries to fashion results into a consistent story of organizational processes (Macal, 2004), and the mathematical physics of uncertainty, which focuses on predicting organizational outcomes. We review the traditional theory of organizations along with our new theory of the mathematical physics of organizations, apply it to two field tests, and then apply the lessons learned to a review of the EMCAS (Electricity Markets Complex Adaptive Systems) agent model designed by the DOE’s Argonne National Laboratory (www.dis.anl.gov/CEEESA/EMCAS.html). Afterwards, we review future research directions and conclude with a discussion of where the results of the traditional model and our model agree and disagree. Methodological individualism, the traditional model of organizations (Nowak & Sigmund, 2004), assumes that interactions between agents generate stable information, I, that an optimum set of perspectives, beliefs, or knowledge, K, subsume other beliefs, and that the aggregate perspectives of individual members is the organization (i.e., an organization is what its members believe). This approach is static (Von Neumann & Morgenstern, 1953), but when repeated generates computational models of dynamics and evolution. The best example is game theory, typically used to promote the advantages of cooperation over competition. Bounding this view, the normative social welfare model of Arrow (1951) indicates that cooperation to achieve a rational consensus as preferences multiply becomes impossible without the imposition of either a dictator or democratic majority rule. From a different direction, Nash’s (1950) possibility of consensus increases as the differences between two cooperating parties sufficiently diminish to compromise over the remainder. In agreement with Arrow and Nash, May (2001, p. 174) concluded that as the heterogeneity of beliefs decreased, a system mathematically becomes easier to control but if it becomes more rigid, it also becomes unstable, requiring a reduction in feedback to maintain control. Thus, in the traditional model of organizations, stability arises from the convergence of consensus-seeking for decisions; a command decision making hierarchy or dictatorship to influence the process; and hierarchical limits to information, I, or censorship. Generally, efficiency increases with centralized command for well-defined solutions to problems (Lawless & Grayson, 2004b). For example, Unilever has concluded that its primary barrier to efficiency is having two CEO’s (www.wallstreetjournal.com, 1/3/2005). However, this method of governance also reduces the organizational forces of innovation, motivating adaptations such as restructures and mergers. In sharp contrast, random exploration among competing organizations and beliefs in democracies generates stable environments and innovation (Fama, 1965; May, 1997). Yet, proving that democracy works so that it can be rationally applied to hierarchical organizations set within democracies or even to computational agent systems is a hard problem. From a review of the mathematics, Luce and Raiffa (1967) concluded that methodological individualism was unlikely to capture the essence of organizations. From a review of group psychology, Levine and Moreland (1998) concluded that aggregating self-reported preference data from the members of a group fails to capture its critical processes. Phenomenologically, Campbell (1996), the founder of social convergence 20 more pages are available in the full version of this document, which may be purchased using the "Add to Cart" button on the product's webpage: www.igi-global.com/chapter/physics-organizationaluncertainty/6791?camid=4v1 This title is available in InfoSci-Books, Business Intelligence, Business-Technology-Solution, InfoSci-Business Technologies, Business, Administration, and Management, InfoSci-Business. Recommend this product to your librarian: www.igi-global.com/e-resources/libraryrecommendation/?id=1