Conceptualization of the use of Artificial Intelligence for Interdependencies Analysis in Requirements Engineering

Abstract

The efficiency in product development is largely determined by the quality of the requirements and the ability of the product design and production planner to analyze them. Interdependencies between multiple requirements identified at an early stage enable a sustainable design of the product as well as the corresponding production system by increasing process efficiency as well as the effectiveness of development processes. However, the necessary analysis of complex interdependencies between requirements of a product and the corresponding production system is time-consuming, error-prone, and highly inefficient when performed manually. Current development processes are based on such manual processes for analyzing requirements in natural language and must therefore be adapted. This paper describes a methodical approach based on a semi-systematic literature review making the complexity of the interdependencies manageable by using existing approaches and methods in the field of model-based systems engineering (MBSE) as well as natural language processing (NLP). Thereby, a transition from informal requirements represented in natural language to analyzable and structured information, which enable interdependencies modeling for requirement chains, is described. A corresponding framework for analyzing interdependencies in the requirements engineering process is derived.