A multi-agent computational model for product development organizations: Integration of distributed AI and organization theory.

摘要

A computational model of organizations emerges as an important new methodology for the analysis, design, and simulation of complex organizations. The key effort of the computational model has been to analyze the dynamic, complex, and distributed nature of organizations, and to design efficient organizations and information processing strategies within the context of specific task environments.;We present a multi-agent computational model for product development organizations. We integrate organizational theories with artificial intelligence methods in modeling product development organizations. In addition, to operationalize the model, we develop an object-oriented simulation system using Visual C++ of MicroSoft.;We evaluate four artificial intelligence search methods for product line design problem and organizational simulation. The search methods compared are: (1) genetic algorithm, (2) tabu search, (3) simulated annealing, and (4) neighborhood search. We investigate how parameters of the search methods affect their performance, and determine the most efficient search method for product line design problems.;In addition, we investigate the impact of the design parameters of product development organizations on the performance of these organizations, such as profits and decision-making time. Finally, we investigate the impact of the external environment of product development organizations on the performance of these organizations.