Optimizing travel costs of feeder-integrated public transport system: A methodology

Abstract

Canada, with a substantial contribution from the personal transport sector, is a major per capita greenhouse gas emitter. This study advocates a sustainable 3-echelon transportation system, integrating Public Transit (PuT) and demand-responsive transit (DRT) for door-to-door service. Electric autonomous DRT vehicles serve the first and third legs of travel, while the second leg relies on PuT. The goal is to identify routes for commuters simultaneously optimizing user, operator, and emission costs. A novel evolutionary algorithm, guided by fuzzy inference systems, optimizes travel costs. The algorithm is calibrated, and its performance is validated against benchmark instances. The proposed optimization framework demonstrates superior performance, achieving quick convergence even for large instances with over 5,000 billion possible routes. Near-optimal routing solutions for sizable scenarios with approximately 100 commuters, 250 PuT nodes, and 50 DRT vehicles can be computed within approximately 20 min.

Keywords

• 3-echelon vehicle routing
• Multi-objective optimization
• Fuzzy inference systems
• Demand-responsive sustainable public transportation
• Evolutionary algorithm