A hybrid knowledge representation based on logical objects

This paper presents a hybrid knowledge representation based on the framework of logical knowledge objects (LKO), which are viewed as abstractions with behaviors, state and inheritance. It accommodates the main features of various knowledge representations, such as rule, frame, semantic network and blackboard, while amalgamating logic programming and object oriented programming paradigms. Therefore, it supports such notions as inheritance, modularity, reusability, structuring and dynamic modification of knowledge, which have turned out to be extremely useful for the design and maintenance of large scale knowledge based systems and expert systems. Using the framework, we can explicitly define the hierarchy of logical objects and easily obtain the solutions. With the basic notions and the semantics of LKO introduced, we focus on how it fits for the various knowledge representations, and discuss the management of knowledge base by exploring the formal specifications of metalevel operations