The importance of mobility and logistics has steadily increased in recent years. Today, this sector represents a significant part of economic output, but also of resource consumption. In order to achieve the ambitious environmental and climate protection goals, there is a high demand for innovative concepts.
In this area we focus on bike and car sharing. They offer considerable potential for saving fossil fuels, particularly in combination with local and long-distance public transport. In recent years, public bicycle rental systems have emerged in many cities around the world. These systems allow automatic rental and return of bicycles at a large number of stations. Modern Car-Sharing providers enable borrowing and returning at almost any parking space within a defined business area (free-floating). We concentrate on the following questions of strategic planning and dimensioning of the systems as well as operations and cooperate with a well-known German free-floating Car-Sharing provider:
Free-floating shared mobility systems offer customers the flexibility to pick up and drop off vehicles at any location within the business area and, thus, have become the most popular type of shared mobility system. However, this flexibility has the drawback that vehicles tend to accumulate at locations with low demand. To counter these imbalances, pricing has proven to be an effective and cost-efficient means. The fact that customers use mobile applications, combined with the fact that providers know the exact location of each vehicle in real-time, provides new opportunities for dynamic pricing.
In this context, we develop a pricing approach for the dynamic online problem of a provider who determines profit-maximizing prices whenever a customer opens the provider’s mobile application to rent a vehicle. Our pricing approach has three distinguishing features: First, it is customer-centric, i.e., it considers the customer’s location as well as disaggregated choice behavior to precisely capture the effect of price and walking distance to the available vehicles on the customer’s propensity to choose a vehicle. Second, our pricing approach is origin-based, i.e., prices are differentiated by location and time of rental start, which reflects the real-world situation where the rental destination is usually unknown. Third, our approach is anticipative and uses a stochastic dynamic program to anticipate the effect of current decisions on future vehicle locations, rentals, and profits. As solution method, we propose a non-parametric value function approximation, which offers several advantages for the application, e.g., historical data can readily be used and main parameters can be pre-computed such that the online pricing problem becomes tractable. Extensive numerical studies, including a case study based on Share Now data, demonstrate that our approach increases profits by up to 13% compared to existing approaches from the literature and other benchmarks.
Shared mobility systems such as car sharing have become a frequently used inner-city mobility option. In particular, free-floating shared mobility systems are experiencing strong growth compared to station-based systems. For both types, many approaches have been proposed to optimize operations, e.g., through pricing and vehicle relocation. To date, however, optimization models for free-floating shared mobility systems have simply adopted key assumptions from station-based models. This refers, in particular, to the part of the optimization model that formalizes how rentals are realized depending on available vehicles and arriving customers, i.e., how supply and demand match. However, this adaption results in a simplification that does not adequately account for the unique characteristics of free-floating systems, leading to overestimated rentals, suboptimal decisions, and lost profits.
In this paper, we address the crucial issue of accurate optimization model formulation for freefloating systems. We formally derive two novel analytical matching functions specifically suited for free-floating system optimization, incorporating additional parameters besides supply and demand, such as customers’ maximum walking distance and zone sizes. We investigate their properties, like their linearizability and the integrability into existing optimization models. An extensive computational study shows that the two functions’ accuracy can be up to 20 times higher than the existing approach. In addition, in a real-world price optimization case study based on data of Share Now, Europe’s largest free-floating car sharing provider, we demonstrate that more profitable pricing decisions are made. Most importantly, our work enables the adaptation of station-based optimization models to free-floating systems.
Over the last decades, shared mobility systems have become an integral part of inner-city mobility. Modern systems allow one-way rentals, i.e. customers can drop off the vehicle at a different location to where they began their trip. A prominent example is car sharing. Indeed, this work was motivated by the insight we gained in collaborating closely with Europe's largest car sharing provider, Share Now. In car sharing, as well as in shared mobility systems in general, pricing optimization has turned out to be a promising means of increasing profit while challenged by limited vehicle supply and asymmetric demand across time and space. Thus, in practice, providers increasingly use minute pricing that is differentiated according to where a rental originates, i.e., considering its location and the time of day. In research, however, such approaches have not been considered yet. In this paper, we therefore introduce the corresponding origin-based differentiated, profit-maximizing pricing problem for shared mobility systems. The problem is to determine spatially and temporally differentiated minute prices, taking network effects on the supply side as well as several practice relevant aspects into account. Based on a deterministic network flow model, we formulate the problem as a mixed-integer linear program and prove that it is NP-hard. For its solution, we propose a temporal decomposition approach based on approximate dynamic programming. The approach integrates a value function approximation to incorporate future profits and account for network effects. Extensive computational experiments demonstrate the benefits of capturing such effects in pricing generally, as well as showing our value function approximation's ability to anticipate them precisely. Further, in a case study based on Share Now data from Florence in Italy, we observe profit increases of around 9% compared to constant uniform minute prices, which are still the de facto industry standard.
Bike sharing has been introduced in many cities, often by municipalities and is nowadays an established alternative for other short-distance transport systems. However, in cities with high elevations, the usual bike-sharing systems face a severe problem. Resulting from an imbalance of demand, the number of bikes at stations at elevated locations decreases during the day, while it increases at stations at lower locations. This situation poses a challenge for the relocation process because high numbers of bicycles have to be transported to the stations at elevated locations in order to achieve a suitable starting point for the next period. With the usage of e-bike sharing-systems, this problem can be circumvented because e-bikes facilitate the mobility in elevated and steep terrains. This paper considers an e-bike sharing-system with removable batteries. In the first step, a deterministic Mixed-Integer Linear Program (MILP) calculates the optimal route for trucks and the optimal initial distribution of bikes. In the second step, a stochastic simulation should evaluate these results.
Airlines today are exposed to high pressure on costs, which will continue to rise as a result of the CO2 tax, for example. This forces them to use their capacities efficiently. The precondition for this is a flight schedule that is as attractive as possible for potential passengers, both in terms of the connections and the flight times. The design of corresponding flight schedules at the medium-term planning level is an extremely complex decision-making problem that can only be successfully tackled with the aid of quantitative methods.
The project focuses on the following two aspects:
Optimizing an airline schedule usually comprises multiple planning stages. These are the choice of flights to offer (schedule design), the assignment of fleets to flight legs (fleet assignment), and the construction of rotations under consideration of maintenance constraints (aircraft maintenance routing). Moreover, the airline must assign crews to all flights (crew scheduling). Traditionally, either these scheduling stages are considered sequentially or an existing schedule is modified to cope with the arising complexity issue. More recently, some authors have developed models that integrate adjacent stages. In this paper, outcomes of a research project with airline information technology provider Lufthansa Systems are presented. We consider the case of a small to medium-sized point-to-point airline with a homogeneous fleet. Hence, fleet assignment is omitted, which offers the possibility to solve schedule design and aircraft maintenance routing simultaneously. Our approach explicitly accounts for passengers’ return flight demand and for marginal revenues declining with increasing seat capacity, hence, anticipating the effects of capacity control in revenue management systems. To solve the arising integrated mixed-integer problem, a branch-and-price approach and a column generation-based heuristic have been developed. An extensive numerical study, using data from a major European airline provided by Lufthansa Systems, shows that the presented approaches yield high-quality solutions to real-world problem instances within a reasonable time.