Agent-based Organizational Cybernetic Approach to Organizational Learning

This paper proposes agent-based organizational cybernetics (AOC), which combines organizational cybernetic framework and computational organization theoretic approach, especially agent-based computational learning model. Organizational cybernetics provides a basic framework in which every organization is comprehensively described to be composed of 4 functional layers: process, coordination, adaptation and self-organization. In organizational cybernetics the difference between the overall purpose of an organization and each function of it plays an essential role especially to provide diagnosis for the organization of keeping viable. In this usage of the organizational cybernetic framework no micro-macro problem between learning of each member and that of the whole system can be seen. On the other hand, computational organization theoretic approach focuses on agents' task resolutions and bottom-up description of organizational activities. Organizational learning process described in computational organization theory mainly consists of error correcting activities by each agent and organization itself based on the agents' decision rules. Computational organization theory describes explicitly models to define detail dynamics on learning in process levels by agents. It is, however, not straightforward to describe double-loop learning as sharing internal models among agents. Though organizational cybernetic approach and computational organization theoretic one have considerably different aspects: on micro-level and on macro-level, we can combine them and build newly emergent model. Our proposed model describes the two loops of organizational learning by representing both processes of learning of internal models and resolving tasks by agents. The model can describe essentially different levels of individual learning and organizational one so as to distinguish effectively the two loops of organizational learning