An Artificial Intelligence Framework for Conflict Mapping and Resolution for Sustainability of Systems

Early design decisions strongly influence environmental, economic and social outcomes, yet sustainability assessment tools rarely reveal trade-offs among these three pillars. This study presents a framework for Conflict Mapping and Resolution for Sustainability of Systems (CONFARM). CONFARM consists of four steps: lifecycle documentation, cause-effect mapping, conflict database construction and multi-criteria scoring. A conflict is recorded when a single decision produces positive and negative effects across pillars. Each effect is evaluated using impact magnitude and pillar weight to generate a sustainability ratio. CONFARM may be applied manually or through automated extraction using natural-language processing and large language models. The method is demonstrated in three sectors representing different data structures and system scales: agriculture (rice and corn), fashion (slow and fast fashion) and energy (nuclear and natural gas). Each system was analysed at increasing conflict densities. Results consistently showed that sustainability scores converged as more conflicts were mapped, indicating stable evaluation across methods. Slow fashion and nuclear systems exhibited relatively higher sustainability performance, while fast fashion and natural gas systems showed lower performance. CONFARM improves early-stage decision support by making trade-offs explicit and enabling comparative evaluation. It offers a structured approach for cleaner production and scalable sustainability assessment across domains.

💡 Research Summary

This paper presents CONFARM (Conflict Mapping and Resolution for Sustainability of Systems), a novel framework designed to systematically identify and evaluate sustainability trade-offs that arise from early-stage design and process decisions. The research addresses a critical gap in current sustainability assessment tools, such as Life Cycle Assessment (LCA), which often fail to make explicit the conflicts and synergies among the three pillars of sustainability: environmental, economic, and social. Early decisions heavily influence lifecycle outcomes, yet the competing effects across these pillars frequently remain hidden until later stages, when changes are costly or impractical.

The CONFARM framework is structured around four sequential steps. First, Product Lifecycle Documentation and Structuring organizes information across five standard lifecycle stages (raw material acquisition, manufacturing, transportation/distribution, use, and end-of-life) using categories like inputs, outputs, and decision nodes. This creates a standardized inventory. Second, Cause-Effect Mapping traces the relationships between design/process decisions (causes) and their resulting sustainability outcomes (effects). A “conflict” is formally recorded when a single decision produces both positive and negative effects across different pillars. Third, identified conflicts are compiled into a structured Conflict Database, enabling traceability and comparison. Each entry details the lifecycle stage, triggering decision, involved pillars, and associated positive/negative impacts. Finally, Multi-Criteria Scoring and Benchmarking quantifies each conflict. Each effect is assigned an Impact Magnitude (based on severity, reversibility, and controllability) and a Pillar Weight (reflecting contextual importance). A Sustainability Ratio is calculated using a weighted sum, allowing for comparative evaluation of different products or systems.

A significant contribution of the work is the proposal for an AI-assisted implementation. The authors suggest using Large Language Models (LLMs) like ChatGPT and Gemini to automate the extraction of conflict relationships from lifecycle documentation texts. This reduces analysis time and enhances the framework’s scalability across diverse domains and data-richness levels.

The methodology is validated through three case studies from distinct sectors: agriculture (rice vs. corn), fashion (slow vs. fast fashion), and energy (nuclear vs. natural gas). These cases represent varying system scales and data structures. The systems were analyzed with increasing “conflict density” (number of mapped conflicts). Results demonstrated that sustainability scores converged as more conflicts were mapped, indicating stable and consistent evaluation. Furthermore, slow fashion and nuclear energy systems showed relatively higher sustainability performance, while fast fashion and natural gas systems scored lower.

In conclusion, CONFARM provides a structured, scalable approach for integrated sustainability assessment. By making cross-pillar trade-offs explicit and enabling both manual and automated analysis, it offers enhanced decision support during early design phases. The framework bridges qualitative cause-effect reasoning with quantitative multi-criteria evaluation, advancing the methodology for cleaner production and holistic sustainability assessment.