Robust integrated vehicle-, crew scheduling and crew rostering in public bus transit

Project: Research

Project participants

Description

This application deals with operational planning in transport companies, consisting of vehicle scheduling, crew scheduling and crew rostering. In vehicle scheduling, the service trips and the operational plan of the vehicles are generated. Such trips are assigned to daily duty schedules in crew scheduling and then assigned to drivers as rosters in crew rostering. These three planning phases are carried out in research and practice sequentially or in better case partially integrated (two of the three phases are considered simultaneously or iteratively coupled). Such a procedure may cause invalid duty schedules or rosters of drivers.At present, there are only a few researches on the integrated operational planning in public transit. A key reason for this is that the longer scheduling period of crew rostering greatly increases the complexity of solving the previous two phases, as it requires monthly resource schedules for integration rather than daily schedules. In our previous research, we have considered the integration of vehicle and crew scheduling as well as different variants of crew rostering individually. Now, the proposed project will examine under which conditions an integration of the three phases of operational planning is solvable for real world problems.Moreover, researchers and practitioners have increasingly interests on robust and efficient plans that allow vehicle plans and duties to remain tolerant to delays and duty cancellations of drivers during the execution. Our existing researched aspects of robust and efficient resource plans will be taken into account in this project and expanded in the new, integrated approach.The project includes three major work packages: modeling, solution approaches, and simulation and pilot studies. In the first work package "modeling", the models (existing / with integrated approach / with and without robustness aspects) are systemized and classified. In addition, benchmark instances are generated for test purposes (artificial and real-world). In the second work package "solution approaches", exact methods for the integrated approach and several alternative approaches (e.g.: specialized heuristics, simplification possibilities of the models / solution methods) are explored. Finally, in the third work package, a simulator will be implemented using real world scenarios to test the efficiency and robustness of the plans generated from different solution approaches. In addition, an empirical study is carried out in cooperation with practice partners and the results are systematized.
StatusFinished
Period01.05.1931.12.22