GLOWB

GLOWB couples AI-accelerated simulators for hydrology, mobility and land use with reinforcement learning (RL). An RL agent searches over policy portfolios — nature-based solutions such as permeable "sponge" streets, bioswales, floodable parks, wetland restoration and shaded walking/cycling spines, combined with demand-management levers like staggered hours, seasonal street reallocation and access pricing — to analyse what to implement, where, and when, subject to budget and maintenance constraints. The centrepiece is an auditable wellbeing reward integrating several components (e.g. undirected travel, daily-needs accessibility, health exposure, mobility performance, environmental quality and built-environment impact). Work is co-designed and evaluated in four contrasting case studies: Odense (compound flood + heat), the Paris–Marne-la-Vallée–Chessy RER-A corridor with RATP (transit-oriented green-social retrofits), Jeonju (monsoon-prone hilly heritage city) and Galway's Salthill–Claddagh waterfront (coastal surge). Municipalities and operators sit inside the consortium, ensuring policies are actionable, embedded in real planning cycles, and transferable.

GLOWB will deliver a wellbeing-centred, NBS-first, AI-accelerated RL planner that learns robust policy portfolios across deep-uncertainty futures and produces interactive, co-produced policy advice for climate adaptation. Concretely: an open policy search engine and simulation stack; validated adaptation portfolios for each of the four testbed cities; and transferable design patterns for other European and international municipalities. Municipal partners in Odense, Jeonju, Paris and Galway will gain evidence-based roadmaps for green–blue infrastructure investments that jointly improve stormwater management, heat resilience, active-mobility uptake and social equity, while safeguarding day-to-day accessibility. Beyond the pilots, GLOWB will shift adaptation planning from heuristic, siloed decision-making toward auditable, multi-objective optimisation — helping cities avoid costly lock-in, reveal blind spots through scenario discovery, and communicate trade-offs transparently to citizens. Outputs feed directly into the DUT Knowledge Hub for uptake across the European urban-transitions community.