Stefan Bock

Station based and free floating are the two established parking regimes of micromobility systems. The former restricts pickup and drop-off to designated stations, which is less convenient for users but reduces the nuisance of sloppily parked rental bikes and scooters. Free-floating micromobility allows users to use any public parking space within the operating area, which increases user flexibility but creates organizational overhead to deal with improperly parked vehicles. Time-varying parking zones, realized via curbside management software (CMS) and geofencing technology, promise a reasonable compromise in the convenience-clutter trade-off. Based on spatiotemporal information, the city administration can either permanently (e.g., pedestrian zones) or temporarily (e.g., weekly farmers’ market) block certain urban areas in their CMS. The micromobility providers, also having access to the CMS, must ensure that vehicles are not returned to undesignated areas with digital fences during the announced times. This paper introduces an optimization approach for micromobility providers to plan time-varying parking zones, given the dynamic municipal parking limitations. Opening and closing urban areas for parking not only requires a digital reaction (i.e., (un)blocking via geofencing) but also produces costs (e.g., removing the remaining scooters of previous periods from the market square). Hence, our optimization task aims to minimize the total costs associated with dynamic parking zones, whereas representative user trips are guaranteed travel within a given time budget. Based on this setting, we show that most urban stakeholders can profit from time-varying parking zones (compared with a static operating area). Our case study based on Berlin-Mitte shows that dynamic parking zones reduce urban space usage at decreasing service costs and only slight user concessions regarding their convenience. Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2024.0855 .

Many online retailers apply scattered (or mixed-shelves) storage in the picking areas of their warehouses. Instead of keeping unit loads together, individual pieces of stock keeping units (SKUs) are stored on various shelves throughout the warehouse. This storage strategy increases the probability that—whatever it is that customers order jointly—somewhere in the warehouse these products will be located in close proximity. Hence, a picker can retrieve them without excessive unproductive walking. The price for this advantage on the picking side, however, is additional effort for the replenishment workers (also denoted as stowers) when restocking the shelves. Instead of moving only a single homogeneous unit load toward an SKU’s designated storage position, each stower has to travel along multiple open shelf spaces until all products on the cart are stored on shelves. The resulting stower routing problem is equivalent to the well-known prize collecting traveling salesman problem (PCTSP). While the PCTSP for general graphs is known to be strongly [ ]-hard, we show that in a warehousing environment, where all open storage positions are located along parallel aisles, it is only binary [ ]-hard. The special parallel-aisle structure allows us to derive an exact solution algorithm with pseudo-polynomial runtime, which solves even instances with hundreds of open storage positions to proven optimality in just a few seconds. Our computational tests show that the performance gains of an optimized stowing process over the status quo, where stowers operate without decision support, are significant. Especially, when the fill level of a warehouse is high, directing stowers on optimized routes promises huge improvements. History: Accepted by Andrea Lodi, Area Editor for Design & Analysis of Algorithms – Discrete. Funding: This work was supported by the German Science Foundation/Deutsche Forschungsgemeinschaft (DFG) by the grant “Routing of human and automated order pickers in modern warehouses” (BO 3148/14-1 and BO 1972/2-1). Supplemental Material: The e-companion is available at https://doi.org/10.1287/ijoc.2022.0173 .