Solving large-scale instances of the urban transit routing problem with a parallel artificial bee colony-hill climbing optimization algorithm

• Published in: Applied soft computing Vol. 167; p. 112335
• Main Authors: Zervas, Alexandros, Iliopoulou, Christina, Tassopoulos, Ioannis, Beligiannis, Grigorios
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.12.2024
• Subjects: Artificial bee colony optimization; Parallel computing; Swarm intelligence; Transit network design; UTRP; Transit network design; Parallel computing; Swarm intelligence; UTRP; Artificial bee colony optimization
• ISSN: 1568-4946
• DOI: 10.1016/j.asoc.2024.112335

Swarm Intelligence simulates the collective behavior of decentralized and self-organized swarms. One of the main relevant methods is the Artificial Bee Colony (ABC) algorithm which simulates the foraging behavior of bee swarms in a colony to produce efficient solutions to various problems. The Urban Transit Routing Problem (UTRP) involves finding an efficient set of routes in a transit network to satisfy travel demand subject to operational and budget constraints. It is a complex, NP-Hard problem, in which otherwise correct solutions can be rejected because of impracticability. In this study, a hybrid algorithm consisting of a parallel ABC and Hill Climbing was used to find quality solutions to the UTRP. Thorough comparative results on Mandl’s well-known instance and Mumford’s large-scale instances demonstrate that the proposed algorithm outperforms existing techniques, achieving high levels of direct trip coverage in small computational times. Remarkably, when applied to the most extensive benchmark comprising over 6 million trips and 60 bus routes, the proposed algorithm demonstrates an impressive 11 % enhancement in direct coverage over the previously best-reported results, allowing the design of real-world bus networks in under 3 hours. •An optimization algorithm based on a Parallel Artificial Bee Colony-Hill Climbing Optimization methodology was developed.•The Artificial Bee Colony-Hill Climbing Optimization algorithm was adjusted to the discrete nature of the problemat hand.•It is the first time that this kind of algorithm is used to efficiently solve large-scale instances of the UTRP.•Algorithm's performance and computational cost were validated with corresponding literature results.