Matching Workflow Contexts for Collaborative New Product Development Task Knowledge Provisioning

Abstract

The variety of product types and spesifications make new product development (NPD) tasks tough work in discrete manufacturing enterprises, which makes it a common strategy to refer to similar outcomes (e.g. the product drawings, work instructions, etc.) of former tasks during NPD processes. To improve the efficiency of discovering similar historical outcomes, this paper presents a novel approach to measure the similarity between task execution contexts in process-aware information systems, and exploit it for runtime task knowledge recommendation. In our framework, the measurement of similarity is preceded by 1) modeling the task context with ontology theory, 2) using the ontology matching algorithms to evaluate the similarities between the corresponding context ontology entities of different tasks instances. The TD-IDF is then utilized to compute the context cohesion between the user’s current task and historical tasks, and the tasks with the highest similarity will be recommended to the task executors, along with their outcomes. Comparative evaluation is performed using TF-IDF, Levenshtein and Affine Gaps, and results demonstrate that the proposed approach achieves good precision and recall, and is efficient in task knowledge recommendation.