Bachelor and Master thesis

Here, you find specific topics tendered by the chair at the moment. If you are interested, please use the “Apply now” button above.

Bachelor theses

Topic: Shelf space assignment for perishable products

Empirical studies show, that assortment size and available shelf space impacts the amount of retailers’ food waste ( see, e.g., Broekmeulen and van Donselaar, 2019a). Customers expect a wide range of products at all time, but especially for slow moving products with short shelf life there is a trade-off between profit from shelf availability and expiration cost. Akkaş (2019) formulates a shelf space selection problem for perishables using an infinite horizon Markov chain model to prevent product expiration and therefore food waste. For this Bachelor thesis, you are asked to

• give an overview on the construction, types and applications of Markov chain models,
• explain the presented model in Akkaş (2019),
• implement the model of Akkaş in Python and solve it with a small dataset, and
• provide suggestions for possible model enhancements and discuss future research directions.

Literature:

• Akkaş, 2019. Shelf Space Selection to Control Product Expiration. Production and Operations Management 28 (9), 2184–2201
• Broekmeulen, R., van Donselaar, K. H., 2019a. Quantifying the potential to improve on food waste, freshness and sales for perishables in supermarkets. International Journal of Production Economics 209, 265–273.

Topic: Food Waste in the Hospitality and Food Service Sector – Methods and Numbers

17% of available food is being thrown away at the retail and consumer level (FAO, 2019), and while a significant proportion of the total food waste generated stems from the household section, the contribution from the non-household section, notably the Hospitality and Food Service (HaFS) Sector, is undeniable. For this Bachelor thesis, you are asked to

• provide a structured review of literature measuring and analyzing food waste in the hospitality sector (e.g., hotels, canteens, restaurants, …),
• present and discuss the numbers with regard to influencing factors (e.g., method, setting, sector, country, …), and
• categorize the methods for gathering food waste data from the literature and discuss their strengths, weaknesses, and applicability.

Master theses

Topic: Markdown optimization for fresh categories (Kaufland)

For a retailer, ordering the right quantity of products is a crucial task on a daily level – too much quantity leads to food waste whereas too little leads to out–of–stock situations, lost sales, and dissatisfied customers. Excess stocks are particularly critical in fresh categories since the items must be sold off within a very short time due to their short best–before date. In order to sell the items before they perish (e.g., before reaching the best before date), targeted discounts are given by the employees in the stores. If a price discount is too high, the remaining stock may be salvaged quickly, and the margin losses may be too high. On the other side, if the price discounts are too low, the risk of food waste increases.

The aim of this master’s thesis is to develop an approach for profit–optimized markdown pricing in fresh categories (bakery, dairy and/ or counter) on a market–specific and item–specific level, with the goal to minimize the total write–offs and waste (losses from price reduction and remaining stock).

For a more detailed description, please follow this link: PDF: Markdown optimization for fresh categories

The tasks include:

• Complete a structured literature review about markdown policies for overstocks
• Structurally derive decision problem and formalize it as a stochastic–dynamic optimization model
• Implement the decision problem as a simulation model (e.g., in Python or AnyLogic)
• Collect data from Kaufland and prepare it for the application
• Complete numerical studies to generate some managerial insights
• Develop plan for potential implementation of the tool at Kaufland

A real–world dataset will be provided

Topic: Predicting Holiday Effects on Retail Sales

For a retailer, ordering the right amount of product is a crucial task on the daily level – too much leads to spoilage/food waste, whereas too little leads to out-of-stock situations, lost sales, and dissatisfied customers. In particular, the days around holidays are of special interest and, at the same time, hard to forecast. The goal of this master’s thesis is to find underlying patterns in customer shopping behavior that help to predict the holiday week’s distribution – even for stores/articles that have never seen this holiday. For this task, a real-world dataset will be provided. On this basis, an exploratory analysis should be performed to identify relevant features to explain differences in holidays and derive probabilities for each day’s share of the week.

For a more detailed description, please follow this link: PDF: Predicting Holiday Effects on Retail Sales.

The specific tasks to be performed include:

• Exhaustive literature review concerning the state-of-the-art concerning the prediction of holiday effects on retail sales
• Data preparation/cleaning of non-holiday effects (second holiday/paycheck day with increased sales within the week, seasonality, external shocks (e.g., Corona pandemic))
• Exploring possibilities to derive holiday week distributions from the data

Topic: Food waste in the hospitality and food service sector

Catering for a large number of people, as is the case in canteens, at events, or in hotels, for example, poses major challenges for service providers. Consumers require an attractive offer of dishes, while their preferences and hunger are hardly known. At the same time, the provider must prevent empty buffets and sold-out meals at all costs, a setting that practically provokes food waste. We want to approach this problem from the providers’ point of view, analyze the current way of planning the quantities and meals in facilities such as hotels, canteens, mass catering, etc. and find strategies for improvement.

The Master thesis should cover the following tasks:

• Literature overview of planning approaches in this area (both empirical and modeling-based)
• Development of an interview guide based on open research questions appearing from the literature review (in coordination with the supervisor)
• Finding and contacting suitable interviewees
• Conducting the interviews and analysis of the gathered information via qualitative content analysis
• Derivation of improvement measures

Topic: The Electric Vehicle Routing Problem

The European Union requires all new vehicles to be zero-emission by 2035 (European Commission, 2021). This leads to a significant change in the mobility sector. At the current point in time, electric cars cannot travel as far and cannot refuel as fast as their combustion engine counterparts.

The Electric Vehicle Routing Problem (E-VRP) extends the classical Vehicle Routing Problem (VRP) with a limited driving range and the necessity of charging a vehicle. The E-VRP for this thesis considers a linear charging function and time windows. The VRP is a generalization of the well-known Travelling Salesman Problem (TSP) to multiple vehicles.

The tasks for students are to:

• Give an overview of the literature on Electric Vehicle Routing Problems
• Implement the mathematical model (E-VRP) of Schneider et al. (2014) in Python and solve it with a commercial solver (Gurobi)
• Implement a column generation algorithm in Python and evaluate its performance against the commercial solver (Gurobi)
• Investigate the effect of varying battery sizes and charging speeds in a computational study

Literature: Schneider, M., Stenger, A., & Goeke, D. (2014). The electric vehicle routing problem with time windows and recharging stations. Transportation Science, 48(4), 500-520.

Topic: The Stochastic Patient-to-Bed Assignment Problem

Patients who are admitted to a hospital as inpatients must be assigned a bed while respecting certain constraints. For instance, gender-mixed rooms are usually to be avoided, and it is desirable to assign patients of similar age to a room since this has been shown to increase patient satisfaction. Transfers of a patient to a different room during their stay are possible but should be minimized in order to reduce work for the nurses and stress for the patients. The goal of the master’s thesis is to go beyond a rolling horizon approach by formulating the patient-to-bed assignment problem with stochastic arrivals and lengths of stay of patients using Markov decision processes (MDPs).

For a more detailed description, please follow this link: PDF: The Stochastic Patient-to-Bed Assignment Problem

The specific tasks to be performed include:

• Generate explicit distributions and test instances from the real data set
• Set up an automated instance generator
• Formulate an MDP for the case with a stochastic length of stay
• Formulate an MDP for the case where both length of stay and patient arrivals are stochastic