Ontology for Knowledge Reuse

为了建立一种构建可复用知识库的方法，我们对本体进行了基础研究，为实现这一目标做出了贡献。本体被粗略地定义为“用作人工系统的构建块的词汇/概念的系统”。它由任务本体和领域本体组成，本文对这两种本体进行了讨论。在任务本体方面，建立了任务本体的描述方案，描述了用于调度任务的任务本体。在领域本体方面，我们设计了流体换热领域中行为和功能的本体及其表示语言。我们的调度任务本体最初是用本体表示语言Ontolangua表示的，但发现我们需要另一种语言。这促使我们设计自己的任务本体表示语言。在此研究期间，我们提出了一种新的本体表示语言和一个本体设计环境CLEPE，CLEPE的详细设计已经完成，目前正在实现中。我们还研究了行为和功能之间的差异，并提出了四个原始维度，称为功能浇头：FT，它们共同表征了两者之间的差异。FBRL：使用FT开发了功能和行为表示语言，并对其性能进行了评估。FBRL用于表示设备的定性模型，这些设备用于通过配置设备来构建系统的仿真模型。此外，它拥有丰富的行为和功能词汇表，使FBRL构建的系统能够轻松地解释它们如何在行为和功能级别上工作。因此，我们开发了一种基于行为和功能本体的语言，并使计算机和人类能够分享对系统如何工作的共同理解。

Aiming at establishment of a methodology for building reusable knowledge bases, we conducted basic research on ontology which contributes to the goal. Ontology is roughly defined as "a system of vocabulary/concepts used as building blocks for artificial systems". It consists of task ontology and domain ontology both of which have been discussed in the research. Concerning task ontology, we established a describing scheme of it and described a task ontology for scheduling tasks. Concerning domain ontology, we designed an ontology and its representation language for behavior and function in the domain of heat transportation using fluid. Our scheduling task ontology was first represented in Ontolingua, an ontology representation language, but found we need another language. This motivated us to design our own language for task ontology representation. Within this research period, we came up with a new ontology representation language together with a environment for ontology design named CLEPE.Detailed design of CLEPE has been finished and is currently under implementation. We also investigated the difference between behavior and function and came up with four primitive dimensions named as Functional Toppings : FT which collectively characterize the difference between the two. FBRL : Function and Behavior Representation Language was developed using FT and its performance was evaluated. FBRL is used to represent qualitative models of devices which are used for building a simulation model of a system by configuring them. Furthermore, it has a rich set of vocabulary of behavior and function which enables the FBRL-built systems to easily produce explanation of how they work at the behavior and function levels. Thus, we developed a language based on behavior and function ontologies and enables computers and humans to share common understanding of how systems work.