Zhiang Lin and Kathleen Carley
Organizations constantly face a dynamic environment where time constraint, external stress, and internal stress all require organizations to respond both quickly and accurately in order to survive (Aldrich 1979; Perrow 1984; Scott 1987). Though the common knowledge is that more complex organizations are more costly, there are debates over whether complexity/redundancy in organizational design would provide organizations with more opportunities to review decisions and whether more opportunities by organizations to review decisions would lead to higher performance. In this paper, we examine the issue: given a set of intelligent agents, how can we design a cost-effective high performance organization. We examine the relationships among organizational cost, opportunities for review, and organizational performance as discussed in the organizational literature by using computational techniques. Organizational performance can be severely affected by time pressure (Lin ,and Carley forthcoming; Means et al. 1992). Organizations must respond quickly if they are to have adequate time to review supporting decisions before making a final decision. Because of the merit of more and fast responses, most organizational researchers have advocated frequent and active responses by organizations (La Porte mid Consolini 1991; Pauchant et ~d. 1990). However, there are several factors that can affect the number of opportunities for review by organizations. One factor is the training. Often, organizations form standard operating procedures in the hope that those programmed procedures will create more opportunities for review (March and Simon 1958).