Research Areas
The department’s current research focuses on the areas of production, logistics, project planning, supply chain management, and marketing. These topics are addressed within the framework of quantitative business administration.
Selected research areas are described below:
1. Dynamic, Centralized Mobility-on-Demand Services
Mobility-on-demand services offer a convenient and flexible compromise between traditional public transportation and private vehicles. There is no need to own a vehicle, and at the same time, users are not dependent on fixed stops and strict schedules. Two well-known problems related to the optimization of mobility-on-demand services are the dial-a-ride problem and the ride-hailing problem. In the dial-a-ride problem, customers can submit a ride request, including their desired time windows, and are then picked up by a type of taxi service. In this scenario, it is possible for multiple customers to share a portion of their route. The customer allows the driver to take a slight detour, and in return, the fare is split among several customers. While in the static Dial-a-Ride problem all ride requests for a planning period are already known at the beginning, in the dynamic variant they arrive over time, once the vehicles are already en route. In the ride-hailing problem, customers want to be picked up immediately. The focus here is less on determining the best possible routes for all vehicles and more on serving pending customer requests as quickly as possible. Typically, significantly more vehicles are available than in the dial-a-ride problem and other vehicle routing problems. This gives rise to two subproblems: Which vehicle should serve which customer request next, and should vehicles without current customers perform anticipatory repositioning to be as well-prepared as possible for future requests? In some problems, ride-sharing is still permitted, while others focus on more traditional taxi services.
2. Location-Routing
Long-term decisions regarding the positioning of buildings, facilities, and equipment within companies can be evaluated and made with the help of location and layout planning. Using appropriate methods, issues such as the positioning of warehouses and distribution centers, the placement of emergency facilities and supply stations, and the determination of an optimal arrangement of machinery in factory buildings can then be resolved. Vehicle routing involves grouping transport or service requests into routes within a specified planning period so that each route can be carried out by a vehicle stationed at a depot. For each route, an optimal sequence of the requests to be carried out must also be determined. Vehicle routing challenges arise, e.g., at freight forwarding companies that deliver and/or pick up goods from various customers, in intra-plant transportation, in waste and hazardous materials disposal, and in winter road maintenance. Optimized location and routing can yield significant synergies in terms of coverage, transit time and capacity utilization, and service quality. Furthermore, a collaborative approach to design and utilize logistics networks by multiple independent companies, combined with the use of various modes of transportation within the framework of multimodal transport, enables a sustainable improvement in the flow of goods and information throughout the entire supply chain.
3. Energy-Cost-Oriented Production Planning
In the field of energy-cost-oriented production planning, the focus is primarily on issues related to scheduling within the medium-term planning horizon. Our objective is to achieve energy-cost-optimal scheduling of all production orders in an order portfolio. The energy costs of an order can be calculated based on the individual electricity requirements for processing on the designated machines, as well as the forecasted electricity prices from the day-ahead market of the EPEX SPOT Germany/Austria power exchange. Scheduling is performed using the forecasted prices, considering either a flow shop or a job shop environment. Appropriate exact and heuristic methods are used to solve the various problem instances.
4. Integrated Production and Logistics Planning
In value-added networks, multiple companies typically collaborate across organizational boundaries to organize, support, and implement procurement, production, transportation, and distribution processes at the operational level using computer-based tools. Since the data volumes generated through systematic information collection or empirical surveys must be processed in a structured manner, it is necessary to apply methods of data analysis, data mining, and forecasting. In addition, the entire planning process should be embedded in a production planning and control system (PPS system) or a powerful Advanced Planning System (APS) should be installed as an extension of a PPS system. However, commercial APS systems do not account for all the necessary organizational and technological specifics in production and logistics, which vary greatly depending on the industry. By using appropriate planning methods, weaknesses can be identified that have not yet been satisfactorily addressed by software. Building on this, production and logistics processes can be accelerated and automated to account for increasing globalization and the dynamic nature of economic life. In this way, costs can be reduced and profits increased in the long term.
5. Resource-Constrained Project Scheduling
Projects can be used to visualize, structure, and reorganize logical and organizational processes. Projects consist of individual activities that consume time and (scarce) resources as they are carried out. In the context of project scheduling, the individual activities must be scheduled in such a way that all time relationships between the activities are satisfied and given resource capacities are not exceeded. Possible objectives of project scheduling include minimizing project duration or ensuring an even utilization of resources over the planning horizon. Many practical problems can be viewed as resource-constrained project scheduling problems, such as complex construction projects, the demolition of buildings or power plants, the development and market launch of new products and software systems, medium-term workforce planning, and the optimization of production processes. Since resource-constrained project scheduling primarily involves human resources, taking behavioral and decision-making patterns into account is highly likely to lead to the successful implementation of projects.
6. Flexible Project Scheduling under Uncertainty
Traditional project scheduling focuses primarily on the static scheduling of deterministic activities within a project. Due to the dynamics and uncertainty of real-world situations, these approaches often fail to account for real-world conditions. Furthermore, the capabilities of employees are often not sufficiently taken into account. Therefore, stochastic activity durations, flexible resource allocation, and varying strengths of resources should be integrated into the models to generate solutions to project scheduling problems that are more easily adaptable to real-world requirements. Use cases can be found, for example, in disaster management. After a disaster, a wide range of activities must be scheduled, and available aid workers must be assigned according to their strengths. This is typically a dynamic situation characterized by significant uncertainties. The resulting project scheduling problems are generally difficult to solve, which is why suitable solution methods must be developed that generate the best possible solutions within a reasonable amount of time.
7. Decentralized, Auction-Based Vehicle Routing
Routes in less-than-truckload (LTL) transportation involve multiple pickup and delivery orders, in which goods must be picked up from one customer and delivered to another within the same route. Various constraints must be taken into account, such as time windows or heterogeneous vehicle fleets of transportation companies. Solving such routing problems can only be done efficiently using heuristic methods,such as a variant of the Grouping Genetic Algorithm. To enable small and medium-sized transportation companies to compete in the road freight market—which is characterized by intense price pressure—entering into a (horizontal) collaboration is a possible solution. This collaboration is characterized by the exchange of as little information as possible regarding the customers of the (competing) cooperation partners, with each partner acting independently. A decentralized, auction-based cooperation mechanism offers a solution to this problem, in which each partner releases specific requests that are offered to the cooperation partners in bundles via a (combinatorial) auction. Following a bidding phase, the requests are distributed among the cooperation partners in a way that maximizes profit within the cooperation. The main challenge lies in bundling the requests in such a way that each offered bundle is attractive to at least one of the cooperation partners, without having knowledge of their customer data. To solve the bundle selection problem, a scenario-based approach can be used, which evaluates each bundle using a metric to generate artificially stochastic bids. In this approach, a bid constellation for each bundle represents a scenario, and the resulting allocation of the bundles constitutes the corresponding solution. Bundles that are frequently allocated across multiple scenarios are then offered to the cooperation partners.
8. Business Applications of Online Survey Systems
Online surveys are the most commonly used survey method in the methodological portfolio of market research firms and are therefore an important tool in market research. This research examines the various applications of online surveys in companies, such as customer or employee surveys, as well as surveys designed to analyze the market environment. This research area focuses on the media-appropriate development of questionnaires for these tools, their implementation using a suitable survey system, the management and execution of surveys, and the analysis and presentation of results tailored to the target audience.
9. Data Analysis in Traffic Safety
In order to improve traffic safety, it is useful to examine the characteristics of traffic accidents—such as alcohol influence, vehicle type, and accident circumstances—over time and to forecast future trends. In doing so, the individual characteristics are not only examined in isolation; rather, interesting combinations of characteristics are automatically identified, and the key geographic regions are pinpointed where appropriate measures should be implemented to prevent specific types of accidents.
All research areas are characterized by a high degree of innovation and technological sophistication, as well as significant practical benefits. In addition to considering monetary objectives, the inclusion of safety and environmental requirements in the objective function gives rise to new challenges that are reflected in the structural properties of the corresponding problems. Quantitative models and solution methods are used to analyze and solve the various problems. During the modeling process, the practical problem is typically transformed into a mixed-integer linear or nonlinear model. Based on the resulting model, optimal solutions can be found for specific problem instances using branch-and-bound and branch-and-cut methods, while suboptimal solutions can be found using priority rule-based methods, local search methods, evolutionary algorithms, and simulation methods. The results obtained are then interpreted, prepared for practical applications, and implemented by suitable project partners in the Hildesheim area.