Keiki TAKADAMA, Takao TERANO, Katsunori SHIMOHARA, Koichi HORI and Shinichi NAKASUKA
Making Organizational Learning Operational: Implication from Learning Classifier System
Abstract:The concepts of organizational learning in organization and management science cover a very wide range of organization-related activities in organization. Since socially situated intelligence is one of such activities, this paper makes the concept of organizational learning operational from the viewpoint of CMOT (Computational ＆ Mathematical Organization Theory) for investigating socially situated intelligence. In particular, this paper focuses on the characteristics of multiagent learning as one of socially situated intelligence, and analyzes them with our model which introduces four operationalized learning mechanisms in organizational learning. This model is a GBML (Genetics-Based Machine Learning) based architecture, and is composed of the following four mechanisms: (a) reinforcement learning, (b) rule generation, (c) rule exchange, and (d) reuse of organizational knowledge. In this model, agents acquire their own appropriate problem solving functions through interaction with other agents in order to complete given problems. A careful investigation on the characteristics of multiagent learning with our model from the viewpoint of socially situated intelligence has revealed the following implications: (1) four learning mechanisms in our model work respectively as (a) a search function, (b) a generator of search methods, (c) an entity to change the search range, and (d) an entity to effectively limit large search ranges; (2) these four mechanisms work effectively by integrating with other mechanisms, in addition to make up for the defects of the other mechanisms. (3) besides the interaction among agents, the interaction among learning mechanisms is required to implement socially situated intelligence at a high level; and (4) there are two levels in the learning mechanisms for multiagent learning (the individual level and organizational level) and each mechanism is divided into two types (single-and double-loop learning). The integration of these various levels and types of learning mechanisms contributes to improving socially situated intelligence.
Keywords: socially situated intelligence, organizational learning, multiagent learning, learning classifier system