The scale of recently developed information systems has been expanded by the increasing complication of system functions. On the other hand, a shorter development period for information systems becomes to be required. The clearance of such a situation becomes important at productivity in information system development and quality improvement. Recently, joint development in distributed environment becomes possible, and, as an idea to improve this situation, we think reusing of the component in object oriented of a programming process is useful. However, reusing of component in the process except a programming one has not been done for the complex relation with the components. In this paper, we propose a design component repository method, in which a design component used in a testing process is stored in a repository and then is mapped onto the products being created during the testing process. Moreover, we applied this method at an actual information system development site to examine its effectiveness. The evaluation results are also reported.

Knowledge management often fails to improve organizational performance under the intraorganizational competition defined by the performance measurement system in an organization. As a reason for this, a harmful performance measurement system leads organization members to their passive stance on sharing knowledge. In this paper we propose an agent-based model of knowledge management under the intraorganizational competition, especially in the case of product-sales organizations. In this model agents learn their attitude for knowledge management for receiving higher marks under the performance measurement system. We confirm the validity of the model by checking the consistency between the stylized facts in the real world and the observed behaviors in the virtual experiment. We consider three effects of initial situations in the organization, environmental contingency, and changes in the weight of the indicators of the performance measurement system. We analyze the three effects on the agents’ attitude and the organization’s behavior under three types of performance measurement system.

Data mining is an analytic process designed to extract useful information from large data sets or databases. The dramatically increase on the volume and complexity of data in modern business and science calls for its emergence and the continuous innovations in computer processing power and data analysis techniques make it possible. Traditionally, data analysis is employed by statistical methods, such as regression, decision tree, and cluster analysis and other methods such as artificial neural network, genetic algorithm etc. Recently, the rough set based method has received considerable attention because it can induce if-then rules from data based on equivalence relation.

In this paper, a rough set based data mining system is developed. A market analysis method is proposed based on rough set theory. Using the developed data mining system, questionnaire data collected from Japanese local real estate market are analyzed. The rules characterizing the relations between the profile of customers and their preference prices of mansion have been obtained, which can help to develop an effective market strategy in the local real estate market.