Article: Collaborative management of dynamic carrying capacity of tourist destinations based on multi-agent deep reinforcement learning and spatio-temporal graph neural networks Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.52 pp.64 - 80 Abstract: This study proposes a collaborative management framework for tourist destination dynamic carrying capacity based on multi-agent deep reinforcement learning (MADRL) and spatio-temporal graph neural network (STGNN). A multi-dimensional topological model is constructed to characterise the spatio-temporal correlation of passenger flow, resources, environment, and service. A STGNN module embedded with spatio-temporal attention is designed to capture dynamic evolution features. A hierarchical MADRL structure realises global coordination. Experiments show that the framework reduces MAE to 0.037, shortens response delay to within 8.2 s, and improves carrying capacity utilisation to 92.6%. It outperforms traditional models in prediction, response, and multi-objective balance, providing an effective method for intelligent and sustainable tourism management. Inderscience Publishers - linking academia, business and industry through research

Title: Collaborative management of dynamic carrying capacity of tourist destinations based on multi-agent deep reinforcement learning and spatio-temporal graph neural networks

Authors: Jingying Li

Addresses: Weinan Normal University, 714099 Weinan, China

Abstract: This study proposes a collaborative management framework for tourist destination dynamic carrying capacity based on multi-agent deep reinforcement learning (MADRL) and spatio-temporal graph neural network (STGNN). A multi-dimensional topological model is constructed to characterise the spatio-temporal correlation of passenger flow, resources, environment, and service. A STGNN module embedded with spatio-temporal attention is designed to capture dynamic evolution features. A hierarchical MADRL structure realises global coordination. Experiments show that the framework reduces MAE to 0.037, shortens response delay to within 8.2 s, and improves carrying capacity utilisation to 92.6%. It outperforms traditional models in prediction, response, and multi-objective balance, providing an effective method for intelligent and sustainable tourism management.

Keywords: tourist destination; dynamic bearing capacity; multi-agent deep reinforcement learning; MADRL; spatio-temporal graph neural network.

DOI: 10.1504/IJICT.2026.153711

International Journal of Information and Communication Technology, 2026 Vol.27 No.52, pp.64 - 80

Received: 06 Jan 2026
Accepted: 27 Feb 2026
Published online: 21 May 2026 *

Title: Collaborative management of dynamic carrying capacity of tourist destinations based on multi-agent deep reinforcement learning and spatio-temporal graph neural networks

Keep up-to-date