Knowledge management in sustainable energy transitions: An exploratory agent-based approach

Sustainable energy transitions pose governance challenges that cannot be addressed through technological optimization alone: they require the management of heterogeneous forms of knowledge — technical, local, tacit, and institutional — distributed among actors with diverging interests and evolving over time. This work proposes an exploratory approach based on Agent-Based Modeling (ABM) as a proof of concept for decision-support architectures oriented toward territorial knowledge management. The application context is the Apulia region, with specific reference to the municipality of Giovinazzo as a conceptual testbed. The model, developed in NetLogo with a 150-month horizon, simulates interactions between strategic and operational agents through two cumulative systemic states (ecosystem health and territorial consensus), interpreted as proxies of tacit environmental knowledge and distributed institutional trust. Feedback flows among agents affecting approval probability represent the main source of non-linearity in the system. Three policy scenarios are compared by varying the operational mode of the regulatory authority: permissive, balanced, and conservative. Scenario A (environmentally permissive) maximizes land conversion but leads to the collapse of both indices, with institutional sensemaking effectively blocked and delayed social response. Scenario B (balanced) keeps both systemic states above critical thresholds through a mechanism that can be interpreted as double-loop learning, without maximizing any single objective. Scenario C (environmentally conservative) preserves ecological and social conditions at the expense of energy coverage, illustrating a mirrored form of selective blindness. A structural sensitivity analysis on three parameters — administrative capacity of the municipal technical office, composition of the landowner base, and institutional permeability to social knowledge — shows that the latter exhibits a non-linear response with an identifiable optimal window, a result that directly informs the priorities of a subsequent participatory phase. The main contribution is the proposal of a framework connecting organizational knowledge management with computational modeling, rather than offering predictive claims, and focusing instead on how the cognitive assumptions underlying decision-making can be made explicit and compared. The framework is intended to be transferable to other sustainability domains where success depends on the ability to govern distributed knowledge under conditions of complexity and uncertainty.