SEER: Sustainability Enhanced Engineering of Software Requirements

TL;DR

SEER introduces a three-stage framework to weave sustainability into requirements engineering from the outset, leveraging natural language models and agentic retrieval-augmented reasoning. It systematically elicites sustainability requirements (SRs) from a domain taxonomy, maps FR/NFR interactions to these SRs, and optimizes requirements to improve SR satisfaction, validated across four diverse PURE dataset case studies. The work demonstrates how LLMs, knowledge graphs, and sentence-transformer fine-tuning can deliver granular SRs and actionable requirement adjustments, with explicit completeness checks to ensure coverage. The framework lays groundwork for scalable, domain-agnostic sustainable software engineering, while acknowledging limitations in taxonomy coverage and model stochasticity, and proposing future expansion to include individual-user perspectives.

Abstract

The rapid expansion of software development has significant environmental, technical, social, and economic impacts. Achieving the United Nations Sustainable Development Goals by 2030 compels developers to adopt sustainable practices. Existing methods mostly offer high-level guidelines, which are time-consuming to implement and rely on team adaptability. Moreover, they focus on design or implementation, while sustainability assessment should start at the requirements engineering phase. In this paper, we introduce SEER, a framework which addresses sustainability concerns in the early software development phase. The framework operates in three stages: (i) it identifies sustainability requirements (SRs) relevant to a specific software product from a general taxonomy; (ii) it evaluates how sustainable system requirements are based on the identified SRs; and (iii) it optimizes system requirements that fail to satisfy any SR. The framework is implemented using the reasoning capabilities of large language models and the agentic RAG (Retrieval Augmented Generation) approach. SEER has been experimented on four software projects from different domains. Results generated using Gemini 2.5 reasoning model demonstrate the effectiveness of the proposed approach in accurately identifying a broad range of sustainability concerns across diverse domains.

Figures (9)

• Figure 1: SR Types within each dimension
• Figure 2: An instance of Taxonomy ourwork
• Figure 3: SEER Framework
• Figure 4: SR Elicitor Component
• Figure 5: Snapshot of Sustainability Standard Document for Canada Housing