Pena Park, located within the Cultural Landscape of Sintra, a UNESCO World Heritage Site in Portugal, is one of the country’s most visited destinations, facing increasing challenges associated with tourism pressure, including visitor congestion and reduced visitor experience quality. The park’s complex topography, dense vegetation, and interconnected circulation network make visitor management particularly demanding, requiring computational tools capable of supporting sustainable heritage management strategies. To address these challenges, this study proposes a spatially explicit GPU-enhanced agent-based model to simulate pedestrian dynamics based on an adapted Social Force Model with group behavior dynamics, describing heterogeneous visitor behaviors, crowd formation, and route selection processes under varying tourism scenarios. The model integrates spatial data derived from OpenStreetMap and is implemented using FLAME GPU 2 to leverage GPU parallelization for large-scale and compu tationally efficient, faster-than-real-time simulations, facilitating iterative scenario testing and creating opportunities for future integration with real-time monitoring systems and digital twin applications. The proposed solution aims to support heritage managers and planners by identifying congestion hotspots, evaluating carrying capacity, and assessing alternative visitor management strategies. A proof-of-concept scenario successfully demonstrated the plausibility and applicability of the proposed model. Performance and stochasticity analyses showed satisfactory computational efficiency and model robustness. Results suggest that GPU-enhanced agent-based modeling constitutes a promising approach for sustainable tourism management in protected cultural and natural landscapes and may be transferable to similar heritage destinations worldwide.

--

Agent-Based Modeling,Tourism,Crowding,Pedestrian,Digital Twin

• Ciências da Computação e da Informação - Ciências Naturais
• Engenharia Eletrotécnica, Eletrónica e Informática - Engenharia e Tecnologia
• Geografia Económica e Social - Ciências Sociais