Research On The CAS-Oriented Ontology Construction And Some Related Key Techniques

With its rapidly development and widely application, information system is from low-level to high-level, from closed to open and cooperation, from simple to complex. It is required the units of system having intelligent and adaptive. This kind of unit is calla agent.One character of Complex Adaptive Systems (CAS) is having some adaptive agent. The core requirement of CAS is the semantics of its data should be represented explicitly, in order to be understood by agents. So semantics plays a very important role in CAS.Compared with other data model or modeling tools, such as Knowledge Organization Systems, Conceptual Model, Database Schema and the object-oriented modeling etc., Ontology is a semantics model and can explicitly present semantic of data, so can be used to knowledge representation and reasoning. Ontology is an information model CAS chooses to use to represent semantic of data.The information system for World EXPO 2010 Shanghai (SHExpo in short) is a CAS will be used to serve SHExpo exhibit, SCAS for short. Using ontology to describe its data is a key technique for implementing the system. The results are called SCAS based ontologies. We use an axiomed system based on first-order logic to represent the ontologies.Constructing SCAS based ontologies belongs to field of Ontology construction Method. The current researches on Ontology Method have made a considerable progress, but because of following limitations, the definitions of ontology are inconsistent, and its mechanism of explaining semantics is not clear, and current ontology methods pay attention to informal stages of ontology construct, but lack of studying on ontological structure and formal stages of ontology construct. So it is difficult to be referred while SCAS based ontologies were constructed:1) The current researches are lack of ontology construction mechanism and its corresponding formal description and arithmetic, so it is difficult to decide ontological categories and their structure. So it is difficult to construct SCAS based ontology.2) Because of lacking upstanding design principles and reusing mechanism of upper ontology, constructing SCAS based ontologies must start from very beginning, repeating labors can not be avoided, and ontological consistence cannot be guaranteed, it is neither in favor of finding initial concepts and to explain explicitly data's semantics.3) It is not enough to describe SCAS based ontologies just depending upon existing relations, such as ISA, parthood and locationhood etc., these relations should be supplemented and expanded, and the mechanisms for restricting and comparing and combining should be studied.4) Lacking of suitable converting mechanism for Web information, it is very inconveniently to model Web information and converting them. So it is necessary to study the corresponding mechanism and implementing prototype.In order to overcome these limitations, combined with some related project, an CAS-oriented ontologies construction principle and its corresponding Ontology Method are described. The author presented following innovative researches on ontology constructiong method:1) Based on abstract principle, an Ontological Delta Mechanism (ODM for short) was proposed, so did its corresponding ontology construction arithmetic. This is a bottom-to-up mechanism used to guide ontology construction. ODM and its corresponding mechanism which author proposed had following characters: The mechanism based on abstract principle and basic characters (made up with attributes) of object, but not on relations among these objects. The mechanism can be used to construct ontology rapidly. Based on the structure, an ontology construction mechanism and its arithmetic were proposed. The formal ontology theory is a universal model, not limited with domain.2) An upper ontology design principle and an ontological reduction mechanism were proposed, in order to explain the semantics of ontology and check consistency of SCAS ontologies and simplify ontology design. Through ontological reduction, Logical errors in ontological design and construction can be effectively avoided. Its values rest with reuse of upper ontology and simplifying ontology design, and it is also convenient to determine the semantics of terms used in the ontology.a) The design principle has following characters: It pays attention to universal concepts and relation among them, the only relation is ISA, so its hiberarchy is clearly defined. It has a formally described and semantic explaining mechanism.b) The ontological reduction mechanism has following characters: Based on entity's classification, it has clear semantics. Because of its clear rules, process and relations, the ontological reduction has feasibility.3) When being applied to SCAS based ontologies construction, in order to correctly to describe the related relations, the componenthood, locationhood, containedhood and orientationhood, and the mechanism for combining and comparing among these relations were researched. SHExpo complex structure is described with these relations. The significance lays in simplifying parthood and expending locationhood, and expending relations. The results can be widely used in layout-related and location-related information description. This work has following characters: Mechanisms for constraining relations and combining relations and comparing relations are presented. A complex structure of SHExpo layout was described with componenthood, locationhood, containedhood and orientationhood, and their axioms were discussed. These relations can be used to describe layout-related structure.4) Based on ontology, semantic Web techniques are explored. In order to meet the requirement of non-ontology system to transform from non-semantics to semantics, an ontology method based on meta-ontology was proposed, it presents a hiberarchy structure. Ontology construction based on the model is a process instantiated lay by lay, and the model can be used to present relations neatly and to model web page and its form-based querying interface. Like database schema, Web page structure and Form-based querying interface have some basic components. These components and the relations among them can be expressed flexibly with the model. The resulting models are well defined, and have clear layers, so can be shared. To sum up, this common technique is a good paradigm for knowledge representing and reusing. The model which author proposed has following characters: It made modeling process clearly, distinguished core ontology from context. It is an instantiation process step by step. Stopping rules between the levels is simple.Compared with other related Ontology Method, this Ontology Method has the following characters:1) Based on objects and relations in a domain, an ontology construct mechanism and its formal mechanisms and constructing arithmetic are studied. 2) The relations should incarnate among leaf classes or among objects of same leaf class was presented.3) Mechanism for semantics description of ontology is studied, and ontological categories of upper ontology are used to explain semantics of ontology.Its significance rests with the proposal of a formal model mechanism and the corresponding arithmetic, and mechanism for explicitly explaining semantics, so it is convenient to construct ontology according to the mechanism. ODM and its arithmetic, a design principle of upper ontology, an ontological reduction mechanism, a description mechanism of complex structure of SHExpo ontology, an ontology method based on meta-ontology which the paper proposed had been partly applied to the researching projects supported by Shanghai Science and Technology Developing Fund.These researches can be used to guide related projects, and have promoted ontology-related study, and have values to be widely applied to. We are sure that these researches have provided a valuable theory basis for related researches and applications in future.