Process Knowledge Graph Construction Method for Process Reuse

Xiuling Li,Shusheng Zhang,Sijia Wang,Bo Huang,Rui Huang

doi:10.1051/jnwpu/20193761174

Abstract

Aiming at the problems of process knowledge reuse and sharing led by the difficulty of unified representation of complex and diverse process knowledge, a process knowledge graph construction method for process reuse is proposed. Firstly, to ensure the accuracy and universality of data schema for process knowledge, the basic schema of process knowledge graph is constructed based on step-nc. Secondly, to improve process knowledge graph basic schema, the process knowledge graph extensive schema is established through process knowledge analysis and process knowledge combination. Meanwhile, to construct the process knowledge graph schema, the existing rules of experience are represented by SWRL language. Moreover, to instantiate the process knowledge graph schema, the process cases are analyzed under the guidance of process knowledge graph schema and the similarity between process cases is computed by latent semantic analysis technology. And then, process cases are transformed into the structured process knowledge graph representation, and process knowledge graph data is obtained. Finally, the process knowledge graph construction application platform is developed to verify the feasibility of the proposed method.

Highlights

Aiming at the problems of process knowledge reuse and sharing led by the difficulty of unified represen⁃ tation of complex and diverse process knowledge, a process knowledge graph construction method for process reuse is proposed
To improve process knowledge graph bas⁃ ic schema, the process knowledge graph extensive schema is established through process knowledge analysis and process knowledge combination
To construct the process knowledge graph schema, the existing rules of experience are represented by SWRL language

Summary

Introduction

工艺知识图谱( process knowledge graph,PKG) 是各类工艺知识在概念结点及关系下的实体扩充和丰富,是一种结构化的语义知识库。工艺知识作为工艺知识图谱的表示对象,其分类直接影响工艺知识图谱的最终形态。 1.1 工艺知识分类间的语义相似性。包括以下 2 种情形: Case1 语义相似度。给定 PKGBS 概念集 E = 识概念; 当 Ei ∩ ej = 0 时,Csim( Ei ,ej) = 0, 此时 PKGBS 未包含解析出的工艺知识概念,则将解析出图 3 为工艺知识合并示例。左侧为工艺知识解析得到的概念关系图;右侧为 PKGBS。对于解析的工艺知识概念,由图 3 可知,概念 workpiece 和 part 共有概念数目为 5, workpiece 包括的概念数目为 12,则利用公式 ( 1) 和 ( 2) 可得, 两者的相似性为 0.417,Asim( part,workpiece)

Results

Conclusion