Digital infrastructures result from individual yet interdependent systems evolving in relation to each other. This paper identifies three processes by which individual systems become embedded into digital infrastructures. The first is a parallel process, whereby systems become embedded independently of each other. The second is a competitive process, whereby systems compete for resources and attention and one system usually thrives while the other system loses importance. The third is a spanning process characterizing a situation of boundary-spanning between distinct parts of a digital infrastructure. The three processes, synthesized into a dynamic model of digital infrastructure embeddedness, offer clarity to the question how digital infrastructures evolve. They also provide insight into the emergence of three forms of digital infrastructures: silofied, regenerated, and unified. Reflecting an interconnection view, our research further facilitates an understanding of infrastructure inertia and its associated consequence. Criticality traps should be avoided by considering the right timing for system replacement in the light of growing embeddedness over time.
This work examines combined vehicle and crew scheduling in public bus transit in the context of both robust and cost-efficient resource allocation. Cost-efficiency in terms of minimal planned costs for resource usage is already addressed in numerous contributions providing increasingly sophisticated optimization approaches that are able to compute highly efficient resource schedules. Unfortunately, during operations the resulting tight resource schedules are generally more susceptible to disruptions causing delays or even cascades of delays. Because of dependencies between scheduled vehicles and drivers, even small delays may propagate through the entire network. However, to date there is only little research assessing schedule robustness against disruptions and delays, affecting both vehicle and crew schedules. We consider the mutual dependencies between vehicle and crew schedules and propose solution approaches that are able to consider the propagation of possible delays. The approaches are used to create cost-efficient resource schedules simultaneously capable to absorb delays and to prevent delay propagation. In particular, we present a fully integrated vehicle and crew scheduling approach. To analyze problem characteristics we investigate to what extent sequential, partially integrated, and integrated vehicle and crew scheduling lead to schedules with different quality. Numerical experiments using real-world problem instances examine the influence of these different scheduling schemes on robustness and cost-efficiency. The results stress the relevance of incorporating possible delay propagation when scheduling resources in (especially urban) public transportation. Compared to pure cost-efficient scheduling, the expected delay propagation can be decreased drastically without additional costs. The online appendix is available at https://doi.org/10.1287/trsc.2017.0757 .
Many public transport companies have recently launched projects testing the operation of electric buses. Progressively, traditional combustion engine buses are being replaced by electric buses. In such cases, some stops on bus lines are equipped with charging technology. Combustion engine buses can operate for an entire day without having to refuel. By contrast, electric buses have considerably shorter ranges and need to recharge their batteries throughout a day. For cost-efficient use of electric buses, charging stations must be located within the road network so that required deadhead trips are as short as possible, but attention must also be paid to construction costs. In contrast to vehicle scheduling, which is a more short-term planning task of public transport companies, location planning of charging stations is a long-term planning problem and requires a simultaneous solving of both optimization problems. Specifically, location planning and vehicle scheduling have to be considered simultaneously in order to open up optimization potentials by comparison to sequential planning, since locations of charging stations directly influence the resulting vehicle rotations. To this purpose, we present a novel solution method for the simultaneous optimization of location planning of charging stations and vehicle scheduling for electric buses in public transport, using variable neighborhood search. By a computational study using real-world public transport data, we show that a simultaneous consideration of both problems is necessary because sequential planning generally leads to either infeasible vehicle rotations or to significant increases in costs. This is especially relevant for public transport companies that start operating electric bus fleets.