Locating carbon neutral mobility hubs using artificial intelligence techniques

Abstract

This research proposes a novel, three-tier AI-based scheme for the allocation of carbon–neutral mobility hubs. Initially, it identified optimal sites using a genetic algorithm, which optimized travel times and achieved a high fitness value of 77,000,000. Second, it involved an Ensemble-based suitability analysis of the pinpointed locations, using factors such as land use mix, densities of population and employment, and proximities of parking, biking, and transit. Each factor is weighted by its carbon emissions contribution, then incorporated into a suitability analysis model, generating scores that guide the final selection of the most suitable mobility hub sites. The final step employs a traffic assignment model to evaluate these sites’ environmental and economic impacts. This includes measuring reductions in vehicle kilometers traveled and calculating other cost savings. Focusing on addressing sustainable development goals 11 and 9, this study leverages advanced techniques to enhance transportation planning policies. The Ensemble model demonstrated strong predictive accuracy, achieving an R-squared of 95% in training and 53% in testing. The identified hubs’ sites reduced daily vehicle travel by 771,074 km, leading to annual savings of 225.5 million USD. This comprehensive approach integrates carbon-focused analyses and post-assessment evaluations, thereby offering a comprehensive framework for sustainable mobility hub planning.