Managing a relational database with intelligent agents

A prototype relational database system was developed that has indexing capability, which threads into data acquisition and analysis programs used by a wide range of researchers. To streamline the user interface and table design, free-formatted table entries were used as descriptors for experiments. This approach potentially could increase data entry errors, compromising system index and retrieval capabilities. A methodology of integrating intelligent agents with the relational database was developed to cleanse and improve the data quality for search and retrieval. An intelligent agent was designed using JACK^™ (Agent Oriented Software Group) and integrated with an Oracle-based relational database. The system was tested by triggering agent corrective measures and was found to improve the quality of the data entries. Wider testing protocols and metrics for assessing its performance are subjects for future studies. This methodology for designing intelligent-based database systems should be useful in developing robust large-scale database systems.