We focus on the aspect of "support", especially, the phase of recognition of problems in Decision Support Systems. First, we show the necessary and sufficient conditions for serial decomposition of a given process system based on GSS(Goal Seeking Systems) approach. Based on the theorem, we secondly construct an algorithm to decompose a given process system as a serial connection of process systems. Lastly, we illustrate the meaning of the algorithm with a simple financial problem. According to the algorithm, we can explicitly show a relationship among variables in a given process system as a serial connection of process systems.

We have studied on the general mechanism of narrative generation system and its application to human creative tasks. In this paper, we apply this narrative generation mechanism to advertising domain. This attempt forms a main part of the development of Marketing/Advertising Integrated Support System (MAISS). In MAISS, we intend to implement the system that consistently supports both marketing process and advertising process and stimulates users' idea formation processes by AI technologies. MAISS generates various narrative structures that represent human life styles and the basic constructions of advertising design from input information (a target segment and a product). This narrative divides into next two types; Proposed Life Scene (PLS) that is the pair of the description Scene (CS) that is a events sequence the PLS was expanded into a narrative structure. Through an experiment and discussion with some users, we could confirm the generality of our narrative generation mechanism and its validity in marketing/advertising processes, therefor our next step is to expand and refine the system for more large-scale experiment.

Adaptation and learning processes of individual perceptions in problematic situation in terms of hypergame are analyzed with Genetic Algorithm. The situation modeled in this paper involves divergence of individual perceptions described by their payoff matrices as players in a hypergame. The model repeats one shot games and improves players' payoff matrices by using Genetic Algorithm. The algorithm suggests the way how the information that each player can obtain should be used to improve the payoff matrices to adapt the given situation. As a result of the experiments in this paper, to get effective improvement of matrices some qualitative information about matrices such as the number of Nash equilibriums is needed as well as the absolute value of the difference between an expected payoff and an obtained one from the result of a one shot game.

Advances in new Information Technologies (IT) and external factors such as globalization and competitive pressure have prompted business organizations to embark on reengineering. However, the evidence of payoffs from such investments can be classified as mixed at best. In this paper, we develop an analytical business value model to assess IT enabled reengineering projects and Kaizen projects. The model is based on the notion of complementarity between IT and organizational variables, and deploys lattice theory and the optimization of “supermodular” functions to derive broad qualitative implications for managing new IT investments and organizational change. The fundamental premise of this complementarity theory is that the organizational payoff is maximized when several factors such as technology, incentives, and organizational structure are changed simultaneously in the “right” directions to move towards an “ideal” configuration. The business value framework for optimizing supermodular functions helps derive more specific and operational models that can be deployed by managers to evaluate reengineering and Kaizen projects.