| Budget Amount *help |
¥2,800,000 (Direct Cost: ¥2,800,000)
Fiscal Year 2004: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2003: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2002: ¥1,000,000 (Direct Cost: ¥1,000,000)
|
|---|
| Research Abstract |
Formulations of evolutionary learning models have been done from the viewpoint of agent-based modeling. In particular two aspects for modeling have been employed : hypergame type model and distributive model.
In the hypergame type model, using tree-person network-type dynamic hypergame, the effectiveness of the proposed learning process model for more general settings of payoffs were systematically investigated.
In the distributive model, a new type of distributive models were proposed, which are based on the levels of single-loop learning and double-loop learning as the basic organizational learning. The proposed model is substantially extended from the former model and a generalization of hypergame situation. The learning of agents' situation perceptions in the double-loop learning process of this model was successfully performed in all of the logically identifiable situations.
As an application of the proposed learning model of organizational learning, the evolutionary and learning process of preference functions in a consumer population was modeled, and analyzed in simulation using genetic algorithm. The model provides a framework of analyzing the adaptation process of the consumer population. As one of the most remarkable results, the characteristics of potentially excellent consumers could be obtained.
The research results were presented at some internal and domestic academic conferences such as North America Association for Computational Social and Organizational Science, Pacific Asia Association for Agent-based Approach in Social Systems Science, Japan Society for Instrument and Control Systems and so on.
|
