Again and again, the industry and trade press reports on companies that are facing up to various logistical innovations and challenges, as can be seen from the following examples: The medical technology manufacturer "Dräger" has recently merged its production and warehouse locations, which has led to various adjustments within the logistics chain (e.g. regarding order picking and the transport of raw materials and finished products) (BVL LOG.mail No. 3 of 19.01.2017). The well-known online retailer amazon is planning to build new logistics centres in order to be able to deliver parcels itself instead of sending them as usual with CEP service providers (BVL LOG.mail no. 3 dated 19.01.2017). In Karlsruhe, as part of the "Kalix" pilot project, the participating shops can have their own purchases delivered to their homes on the same day (BVL Magazine "Zwei" 2016). The Swiss company Imaginecargo wants to reduce CO2 emissions with so-called "Bike-Bahn-Bike-Zustellung" (BVL Magazine "Zwei" 2016). And the automotive industry is developing integrated logistics and mobility concepts such as the "Connected Car", in which vehicles automatically exchange information on current traffic flows, traffic light phases, construction sites or free parking spaces (BVL Magazine "Eins" 2016). Behind most of these ideas, concepts and technical developments are quantitative models for decision support. Within the framework of various project, bachelor or master theses, research is to be carried out, presented in a structured manner and, if possible, demonstrated within the framework of a separate example, which quantitative models or procedures are or could be applied in the individual cases.

Many practical problems can be represented by mixed integer linear models, since some variables in reality can only assume integer values, e.g. number of packets to be sent, number of machines to be used, etc. In contrast to linear optimization, the solvability of mixed integer problems is made more difficult by the additional limitation of the decision variables. Exact methods, such as the branch-and-bound method, usually require too much computation to solve problems with a large number of integer variables and/or constraints. For this reason, heuristics are often used to provide a good (but not necessarily optimal) solution within a reasonable time. The aim of the work is to identify the advantages and disadvantages of heuristic and exact solution methods and to implement suitable solution methods for a selected problem in Python or MATLAB. (Note: Basic knowledge of programming is helpful)

Natural disasters are characterised by extreme uncertainty, as neither the time nor the place nor the extent of the destruction are known before they occur. However, in order to be able to provide the aid organisation with the best possible support in preparing for a natural disaster, models must be developed that take account of uncertain input data. The aim of the work is to present different modelling approaches for the consideration of uncertainties and to compare them with each other. Building on this, mathematical models are to be examined for the following questions, among others: Which uncertainties make the decision-making of aid organisations particularly difficult? Which modelling approaches are particularly frequently used in the literature on disaster logistics and how suitable are they for recording real uncertainties?

Smooth logistics are essential to ensure the success of a humanitarian operation. The requirements depend on the type of disaster and the phase the operation is in. In this work, a specific situation from humanitarian logistics should be selected and modelled (e.g. the reaction to an earthquake or the preparation for a tsunami). First, the literature will be examined for existing modelling approaches and typical features of this situation and possible modelling approaches will be worked out. Building on this, an own modelling approach will be developed. This will then be evaluated with the help of one or more suitable scenarios.

For each type of new Mobility Service, there are different questions that can be answered with the help of optimization methods.
In the case of stationary or free-floating bike sharing, the routes of the service transporters for replacing and repairing the bicycles or the question of optimal new locations for bicycle stations are topics that can be mathematically optimized.
Car sharing offers similar research topics to bike sharing. An additional question is, for example, the optimization of the refueling (with fuel or electricity) of the vehicles in terms of time and/or location.
Ridesharing, which includes ridehailing (taxi rides), carpooling (commuter communities) and ridepooling (collective taxi rides), can benefit in different ways from methods of Operations Research. For example, vehicle routing and matching between vehicles and customers can be optimized, or dynamic pricing can be used to maximize revenue. On the Uber Engineering Blog you can find some articles and papers on the topic, e.g. "Dynamic Pricing and Matching in Ride-Hailing Platforms".
The probably most recent mobility service in Germany is E-Scooter sharing. E-Scooters are electrically driven scooters. Scientific literature on the optimization of this mobility service is still scarce. A research object is, for example, the optimization of evening rides, which are necessary to collect the scooters and charge them over night. E-Scooter sharing providers include myTaxi with the brand Hive, Bird, Lime or Tier Mobility.
The implementation of mathematical models, which is usually part of every thesis, and/or the analysis and evaluation of data can be done in GUSEK or Python.
The concrete topic is to be developed in consultation with the supervisor.

Operations Research is focused on the formulation of mathematical models to simplify real-world planning problems and on the use of algorithms for solving these models. For more than 20 years research has been conducted on the BAP and a variety of model formulations were developed and improved. For a better understanding of the underlying concepts, model clusters are to be identified for the different model formulations within the relevant literature.
The main goal of your thesis will be the identification of basic mathematical models representing the model formulations from the relevant literature. Therefore, you will (1) provide an overview of the relevant literature, (2) group similar models based on their model formulation and (3) present the general, basic model formulation of each group. Moreover, you can apply the basic models of each cluster on a case study and analyze, e.g., the necessary computational effort and the objective function values.
Your results will feed into our research on the optimization of berth allocation scheduling at container terminals. The implementation of mathematical models, which is usually part of every thesis, and/or the analysis and evaluation of data can be carried out in Python.

As BAP is a real-world problem, where information may be incomplete or incorrect, the actual vessel arrival time and the handling time are uncertain or unknown when the vessel’s berth is to be scheduled. Reactive approaches, e.g. disruption management, aim at maintaining the berthing schedule’s feasibility, when newly gathered information differs from the (original) expectations, i.e. a disruption occurs.
The main goal of your thesis will be the application and analysis of a reactive optimization model for the BAP. Furthermore, the basics of the BAP and current literature on the BAP under uncertainty are to be presented and discussed.
Your results will feed into our research on the optimization of berth allocation scheduling at container terminals. The implementation of mathematical models, which is usually part of every thesis, and/or the analysis and evaluation of data can be carried out in Python.