A public transit company’s planning process generally involves many steps that include: network design, frequency setting, timetabling, vehicle scheduling, crew scheduling, and rostering. To reduce complexity, these tasks are often tackled sequentially, with each step providing the starting point for the next. However, it is possible to obtain significant savings when multiple steps are combined and optimized simultaneously. A previous paper (Integrated vehicle and crew scheduling in practice) explained vehicle and crew scheduling can be combined. This paper discusses the integration of timetabling and vehicle scheduling, and focuses on the main issues that must be addressed in order to reach viable solutions, as implemented in the HASTUS-NetPlan module...
Integrated vehicle and crew scheduling in practice
Public transit providers are facing continuous pressure to improve service quality and reduce operating costs. Flexible and powerful optimization algorithms have thus been developed and used for many years to help them with this challenge. Vehicle and crew scheduling are among the most studied problems in this area. These problems can be solved independently (in a sequential manner) or simultaneously (integrated vehicle and crew scheduling). In the latter case, models become more complex and grow in size, but may lead to significant savings if they can be tackled efficiently. Among the research groups that have recently contributed to this topic, we note GERAD, ECOPT, and ZIB. The aim of this paper is to report on applications of integrated vehicle and crew scheduling with several transit companies worldwide, and discuss our experiences and some of the practical issues we have encountered...
Transit timetable synchronization: Evaluation and optimization
Trip synchronization is an important issue for most public transit companies. Minimizing waiting times for transfers between trips can improve customer satisfaction, which in turn leads to increases in ridership and revenues. Several papers have been published on this subject in recent years. In most cases, researchers have studied straightforward mathematical formulations designed to generate timetables with maximal synchronization between trips and have developed heuristic optimization methods for the non-linear models that arise in this context. Although some models consider other factors such as headway regularization and trip frequencies, very few address more complex issues such as the impact of these timetables on the overall vehicle scheduling process...
Dynamic constraint generation in CrewOpt, a column generation approach for transit crew scheduling
Transit crew scheduling is a challenging practical optimization problem that has stimulated a lot of research activity over the past three decades. Although several algorithmic approaches to this problem have been tried over the years, column generation is generally considered the most powerful method currently available. First used commercially in Public Transport by GIRO, column generation was initially restricted to problems of small to moderate size. Since then, significant developments have enabled it to solve problems of all sizes efficiently, leading GIRO to select CrewOpt as the standard crew scheduling algorithm for all our clients...
