In a genetic algorithm (GA), the undesirable phenomenon of excess convergence can often occur. Excess convergence is the phenomenon where the diversity of a group is lost. This phenomenon occurs because homogeneous individuals are increased rapidly in the group while searching, Therefore, a crossover loses its function. Once the excess convergence occurs, the search by the GA becomes meaningless. It then becomes important in the adaptation of the GA to avoid this excess convergence and maintain the diversity. First, we show an implementation of a parallel GA based on a multiple group type island model by using object-shared space. Next, as an effective method for avoiding the excess convergence in a simple mechanism, we propose the diversity maintenance technique based on selection of the homogeneous individuals called Noah's ark strategy in parallel GA, and clarify the effectiveness on the knapsack problem. Our proposed method is to replace the individuals of sub-groups having the excess convergence with the new individual coming from the search space. That is avoid the excess convergence by expelling homogeneous individuals asynchronously except for Noah who is an elite individual, and limiting a decrease of the diversity of the entire sub-group.

View full abstract

　Among the researchers who investigate the society in Heian and Kamakura periods, a comprehensive database of the persons and literatures together with a user-friendly retrieving service has been required. In this paper, we present a database system based on Buddhist scriptures and genealogical trees of both the blood relationship and the master-pupil relationship. We experimentally construct the database, while we implement the retrieval services which apply to the database. With these services, the user can not only see a single genealogical tree but obtain two genealogical trees to understand the human relationship between the specified two persons. Moreover we consider how to store a number of records of person and literature smoothly in the database. Finally, we propose and implement a constructing method of the genealogical trees by a bottom-up approach. By using this system, the researchers can register the records efficiently, and they will find novel human relationships behind the genealogical trees.

View full abstract

　The research on text of the "Tale of Genji" is still imperfect. Much manuscripts which is not known yet exists on the earth. We have to conduct those investigations. However, we hardly argued about the method of arranging the data of investigated texts. This paper reports the data management by the Comma Separated Value. And processing of the CSV file by the tool "kogetsu" and "takai" is reported. "Kogetsu" is the tool which collates different texts. "Kakai" is the tool which measures the distance of texts by comparing many texts. When the data of much manuscripts are accumulated, such a processing tool will bear a big role.

View full abstract

　We use Macroscopic Anomaly as one of the methods of trying earthquake prediction at this laboratory of Okayama University of Science. Generally, Macroscopic Anomaly just befor the occurrence of an earthquake is observed in the circumference of the focus. In order to use this macroscopic anomaly and to perform an earthquake prediction, a quicker collection and public presentation system is required. For this reason. MASPA (Macroscopic Anomaly System by c-PISCO ASP) were started from September, 2000 on the website of a laboratory, Functional strengthening of MASPA and the employment ersult by December, 2002 is introduced.

View full abstract