A semantic framework for resilience analytics and AI-assisted decision support in growing and changing socio-technical systems

Abstract Socio-technical systems are becoming increasingly interconnected, heterogeneous, and dynamic while continuously evolving through new technologies, infrastructures, policies, resources, and operational practices. At the same time, critical infrastructure operators, businesses, and public authorities face growing challenges arising from climate change, cyber threats, geopolitical instability, supply-chain disruptions, and evolving regulatory frameworks. Supporting resilience and sustainability under these conditions requires analytical capabilities that extend beyond static digital twins, application-specific simulation models, and fixed system representations. This paper presents FRAMESS (Framework for Systemic Risk Analysis and Exploring Sustainable Solutions), a semantic and extensible framework for resilience analytics, sustainability assessment, and AI-assisted decision support in growing and changing socio-technical systems. FRAMESS enables the continuous development, integration, simulation, and evaluation of evolving system models despite distributed modelling expertise, changing system architectures, emerging hazards, and large combinatorial design spaces. The framework combines semantic system modelling, flexible system models (FSMs), multi-agent simulation, scenario and disruption management, integrated assessment methods, and AI-assisted exploration within a unified analytical environment. A central contribution is the semantic representation of agents, resources, dependencies, subprocesses, hazards, and adaptation measures, enabling interdisciplinary model integration while preserving extensibility, interoperability, and analytical consistency. The framework further supports resilience quantification and benchmarking by enabling systematic comparison of alternative system configurations under multiple hazard conditions. The capabilities of FRAMESS are demonstrated through a resilience-oriented planning application for allocating strategic reserves in a food supply chain under climate change-induced drought scenarios. The demonstrator illustrates how the framework supports resilience assessment, evaluation of adaptation strategies, and AI-assisted exploration of alternative system configurations. The paper argues that future resilience management requires semantic and extensible analytical environments that integrate evolving digital twins, resilience analytics, and AI-assisted exploration to support resilient and sustainable transformation pathways under uncertainty.