| Abstract | During the last two decades, Logic Programming (LP) stands as a separate paradigm and becomes more popular. At the beginning, applications of LP were limited to mechanical theorem proving. Now, many other applications demand logical paradigm in their domains, e.g., Artificial Intelligence (AI), databases, knowledge bases and distributed systems. Many such applications, however, need more expressive power. These applications are mainly characterized by two kinds of knowledge: Static and Dynamic. Conventional logic provides a suitable framework to represent static knowledge. Due to the lack of a unified framework to represent both kinds of knowledge, many extended knowledge representation systems have been proposed. Some logical formalisms have addressed the problem of a unified framework without referring to the dynamism, and some have addressed the problem of dynamism in LP without referring to a unified framework. So far, the notion of dynamism and a unified framework for logic programs have been dealt separately. The purpose of this research is to link these two fields together by providing a unified theoretical foundation for dynamic knowledge representation systems. First, a theory of updates is axiomatized at an abstract level defining basic structures, called transaction representation systems and transaction parameterized representation systems. Logic programs are developed on transaction parameterized representation systems, called Transaction Generalized Logic Programs (TGLPs). One of the most important features of TGLPs is that they are defined not by syntactic characterization but by axiomatic characterization. Next, the model theoretic semantics and least fixpoint semantics are defined formally without referring to the concept of variables. Besides, a path specialization system is developed for constructing theory for dynamic knowledge representation systems where both declarative and procedural semantics of TGLPs can be discussed. Then, a proof theory is developed for TGLPs by providing a more pragmatic and elegant set of inference rules. Two logical formalisms have been developed under the TGLPs theory. First, Transaction Logic (TR) is defined as a special case of TGLP and later a framework for multimedia databases modeling is developed. The major result of this research includes the development of a unified theoretical foundation for dynamic knowledge representation systems which provides declarative semantics and a set of sound inference rules. |