Accession Number:

ADA339606

Title:

Computational Models of Human Organization Dynamics

Descriptive Note:

Quarterly status rept. 16 Oct 97-15 Jan 98

Corporate Author:

SYNERGIA REDWOOD CITY CA

Personal Author(s):

Report Date:

1998-01-15

Pagination or Media Count:

5.0

Abstract:

During this project period we have made significant progress on the development of the mapping, modeling, and analysis technologies we described in our previous quarterly report. We have also developed an agent architecture, with respect to which agent practices have their operation. This architecture defines how agents receive information from the world, how they select practices for execution, how they resolve ambiguity over which practices to undertake i.e., resolve conflicts cope with uncertainty, and how they execute practices. We have also developed a complete and realistic crisis scenario a Noncombatant Evacuation Operation NEO occurring in Tanzania. The scenario has been reviewed and correctedenhanced by special operations, army planning, and Africa experts. This scenario has a wealth of intentional actors whose practices are interdependent and conflicting. Some of the intentions and even some of the actors are hidden. The latter leads to anomalies conflicts between trusted data and implications of a trusted model. Anomalies are captured as part of the validation studies aimed at testing and verifying the emerging organization model. We show that anomalies can be used as the basis for real time critique and revision of models. We thus illustrate that organizational modeling, model critique and validation, and analysis can occur not only prior to but also during operations. In other words, the tools are suitable for in situ analysis e.g., planning based on simulation of organizational dynamics under various threat scenarios, in addition to their role in off-line studies of organizational practices e.g., aimed at learning and redesign.

Subject Categories:

  • Psychology

Distribution Statement:

APPROVED FOR PUBLIC RELEASE