In the context of a very large knowledge base which consists of rules and the extensional database of facts, the major problem is how to process the relational operations on the extensional database in the minimum amount of time to satisfy the user queries. To speed up the relational operations on very large knowledge bases, parallel processing is essential. But most of the previous index schemes and file structures were designed for sequential processing environment, so they can not be easily mapped to parallel processing systems. One reasonable indexing scheme for parallel processing systems is the concatenated code word (CCW) surrogate file which has small size and simple maintenance requirement. The CCW surrogate file can be used to index the rules as well as facts, and relational operations can be directly performed on the CCW surrogate files. CCW surrogate files can be mapped well into the parallel architectures because their structure is quite compact and regular. The main goal of this research work is to develop optimal clustered CCW surrogate file and data file structures suitable for parallel processing systems. Parallel relational operation algorithms based on the clustered file structures that are designed in this project will be evaluated and their performance compared with other knowledge base management systems. The efficient index and file structures developed for very large knowledge bases support parallel relational operations used in logical inference mechanisms, thus making these systems more effective.
