Semantic Query Optimization in an Object-Oriented Semantic Association Model (OSAM)

Abstract : The Objective Semantic Association Model (OSAM*) represents the semantics of a knowledge base in terms of structural properties, operational characteristics, and knowledge rules associated with the objects of concern in an application. Because of its rich semantic modeling capabilities, OSAM* can be used as the underlying model for constructing a more powerful knowledge base management system (KBMS) than those management systems that use the traditional relational data models. Knowledge rules in a KBMS serve to uphold the semantic properties and security/integrity constraints and to derive data that is not explicity stored in the knowledge base. The knowledge rules in OSAM*'s KBMS can be specified in two ways: (1) frequent integrity constraints are structurally specified by key-words in OSAM*'s Semantic-diagram (S-diagram) and (2) user constraints and deductive rules are specified by a rule specification language. The knowledge rules are important sources of information for query optimization (i.e., semantic query optimization). This thesis (1) analyzes the integrity constraints and deductive rules (knowledge rules) allowable in OSAM*, (2) defines a set of transformation rules which can be used for query optimization, (3) uses these transformation rules in a number of case studies to show how query plans can be generated, and (4) discusses the cost estimation for evaluating the query plans. Keywords: Theses high level languages.