AI-driven eco-routing and tourist behavior modeling for sustainable destination marketing
Open Access
- 22 September 2026
- journal article
- Published by LLC CPC Business Perspectives in Innovative Marketing
- Vol. 22 (4) , 62–82-82
Abstract
Type of the article: Research Article Abstract The growing pressure on tourism systems caused by overtourism, environmental degradation, and uneven spatial distribu-tion of tourist flows highlights the need for adaptive and sustainability-oriented management approaches. This study develops and tests an AI-driven eco-routing and tourist behavior modelling framework (ER-TBM) as a choice-architecture tool for sustainable destination marketing. The framework combines a fixed graph representation of a synthetic destination, Markov decision states, tabular Q-learning, and congestion-sensitive recommendations. To address reproducibility, the benchmark specifies ten points of interest, 22 undirected edges, all edge attributes, tourist preference and fatigue distributions, Poisson arrivals, recommendation compliance, congestion and emission functions, training settings, stopping rules, and random seeds. Five baselines and ER-TBM were evaluated under normal, peak, and event-demand conditions using 30 paired independent runs per scenario with 120, 180, and 240 agents, respectively. Relative to Dijkstra routing, ER-TBM reduced mean congestion exposure by 5.1-5.3% and increased tourist utility by 0.6-0.7%, while mean CO2 remained statistically indistinguishable and monetary route cost was 9.4-9.8% higher. The simulation-based Tourism Sustainability Index ranged from 0.712 to 0.732 for ER-TBM and did not show practically meaningful superiority over optimized baselines. The findings indicate that behavior-aware eco-routing can redistribute simulated visitor flows without a statistically significant increase in emissions, although this benefit is accompanied by higher monetary route costs. From a sustainable destination marketing perspective, the proposed framework represents a computational choice-architecture proposition. Its potential to influence actual tourist behavior, recommendation acceptance, destination image, satisfaction, and revisit intention requires further field-based validation.AcknowledgmentThe research was conducted within the framework of the prospective development plan for the scientific field “Social Sci-ences” No. BF/25-2026 on the topic “Formation of a Competitive Development Model for the Ukrainian Economy: Human Capital, Financial Stability, and the Development of Innovative Industries amid War-Related Challenges and the Emergence of the Sixth Technological Paradigm”.Keywords
This publication has 34 references indexed in Scilit:
- (Big) Data Driven Strategic Decision-Making in Overtourism Management2025
- Smart Tourism2024
- Graph neural networks: A review of methods and applicationsAI Open, 2020
- Exploring the impact of traffic congestion on CO2 emissions in freight distribution networksLogistics Research, 2016
- Multidestination trip, tourism2015
- Smart tourism: foundations and developmentsElectronic Markets, 2015
- Monitoring and Managing Visitors Flows in Destinations using Aggregative GPS Data2010
- Tracking tourists in the digital ageAnnals of Tourism Research, 2007
- Modeling Tourist MovementsAnnals of Tourism Research, 2006
- Markov decision processesEuropean Journal of Operational Research, 1989