Example projects BSc thesis

Here you can find examples of BSc thesis projects of the Operations Research and Logistics group from the recent past. These example projects are to be regarded in terms of themes within ORL in general; you can also check these examples in case you are looking for inspiration for an MSc thesis.

The aim of the examples is to give inspiration about possible thesis projects at ORL, and about future work with a specialization in ORL. Unfold the items below to read more about the presented titles.

If you want to discuss your options for a thesis at ORL, you can contact us via: education.orl@wur.nl. For more information about prerequisites and procedures, see the ORL thesis and internship page.

BSc thesis projects

The link between inventory and service levels in supermarkets (WU)

This BSc thesis is about the inventory in the supply chain, service levels and an application of these concepts on supermarkets. The different types of inventory are decoupling inventory, cycle inventory and safety inventory. Which type a company has depends on their strategy. Almost all companies work with service levels. A service level is a measurement of product availability for customers. There are two types discussed: The α-service level and the β-service level. The α-service level is also called the cycle service level (CSL) and the β-service level is also mentioned as the fill rate. For both types there are several formulas to calculate the level of inventory and the order-up-to level with a given service level.

The level of inventory in supermarkets can be influenced by several factors. The first factor is the influence of the assortment of a supermarket. The second factor is the fact that supermarkets have promotion activities. The practical side of inventory management in a supermarket is compared with the theory.

Optimality of Resource Distribution. An application for improved sustainability of food supply chains. (TIFN)

In the context of supply chain sustainability improvement, Multi-Criteria Decision Making is a usual tool to determine an optimal resource distribution. However, applying this type of analysis presents a number of practical inconveniences. To avert these drawbacks, the study tries to adapt a different approach, originally introduced in healthcare, to supply chains. An analysis is made of this alternative programming model, as well as an explanation of the adjustments required to convert the model from the healthcare to the supply chain context.

Understanding and finding the applicability of a specific multi-criteria approach for lot sizing problems

Multi-criteria decision making models aim to support decision makers in taking balanced decisions if several, mostly conflicting, goals are involved. The goal of this research is to understand a specific multi-criteria model from literature and to apply it to lot sizing problems. In order to solve the models the mathematical programming package Xpress-IVE has been used. The research started with understanding the main differences between the contradictory principles ‘Utilitarianism’, ‘Egalitarianism’ and ‘Rawlsian’. The main distinction  between Egalitarianism and Rawlsian refers to  the concept of ‘Pareto Optimality’. In contrast to Egalitarianism, the Rawlsian principle always searches for Pareto Optimal solutions. Therefore, it is decided to combine Rawlsian (and not Egalitarianism) with Utilitarianism in a single model. The research revealed that  combining both principles in a single model is only possible if the Utilitarianism objective and the Rawlsian objective share the same unit. Otherwise, both principles cannot be compared with each other. Next, the specific modeling approach is applied to a simple lot sizing problem, which aims to find balanced solutions for two conflicting objectives terms, i.e. set-up costs versus inventory holding costs. To conclude, the main result of this research is that there is no possibility yet to apply the specific modeling approach to lot sizing problems. New reformulation are needed for the applicability of both principles to lot sizing problems.

Vehicle routing problems with additional address information

The e-commerce market is increasing, more and more packages are ordered online and they need to be delivered. Nowadays courier companies deliver packages at one specific address. But could there be a more efficient way to deliver packages in case more addresses are linked to an order? The aim of this thesis was not only a literature research but also to develop and test models for the multiple adress problem. A brief overview of existing literature on routing problems is given. In an hierarchical structure a distinction is made between the highest level of routing problems, the categories of routing, and the problem classes of one particular problem category. Routing problems, i.e. vehicle routing problems (VRP) and the (multiple) traveling salesman problems ((M)TSP) are some of the discussed problems. For solving a TSP two approaches are distinguished. The first approach aims to find the optimal solution and the second one are methods based on heuristics. In this thesis a mathematical programming based heuristic is developed for adding additional addresses to customers.

Methods for generating near-optimal solutions for a single-objective diet model

Due to the growth of the  world population and the increasing number of people that struggle with their weight, there is a need for dietary guidance and diet models can help with that. Whereas these models normally result in the generation of an optimal solution, near-optimal solutions might be better than the optimal one in some cases. One could think of the sake of variety as well as of the real world constraints or preferences that cannot be modeled. Two methods already exist for generating near-optimal solutions for single objective continuous linear programming problems: the Hop-Skip-Jump method and the Random method. This paper presents two new methods: the MaxDistance method and the MaxMin method.  Of these two methods, the MaxMin method performs best in generating near-optimal solutions.

Demand- and supply chain uncertainties

The landscape for buying in supermarkets is rapidly changing. A few years ago, just some of the supermarkets in the Netherlands had their own website, nowadays supermarket have home delivery systems, and online retailers use drones to deliver their products. This rapid change goes hand in hand with increasing uncertainty. This uncertainty can cause effects in demand- and supply chains(DSC). In this research thesis four DSC are identified. This thesis developed a framework to compare these different DSC with each other on the basis of Key Performance Indicators to mutually compare DSC. The framework provided can also be used for further research in this subject. In addition, the four DSC are scored in the context of uncertainty to see how these DSC perform on the bases of key performance indicators or parameters and what the differences are. The main question of this research thesis is to improve customer satisfaction by means of DSC uncertainties. This research found that most uncertainty in the researched DSC can be managed through advanced Information Systems, and it is important to pursuit the idea of collaborating DSC to manage uncertainty and logistical problems. ​

How to support the construction of tours for parcel delivery with multiple locations?

Due to the popularity of webshops, the industry of delivering parcels  expanded substantially. Consumers are more and more in control and become even more demanding. In order to meet the consumer’s rising expectations; delivery services need to reduce their costs by delivering as much and as quickly as possible. To accomplish this goal the routing of couriers needs drastic adjustments. If the consumers could give multiple addresses for parcel delivery, the routing problem may become more effective for the parcel industry. By using literature and professional software it is examined whether the concept is actually beneficial. A framework has been created to give an overview of what types of routing models exist, when to use it and how to apply it to the specific case. The proposed framework shows that the concept of multiple locations for one parcel is a traveling salesman problem. By modeling and understanding the basic traveling salesman problem, it gives insight in how to adapt the model with a single addresses into a model for multiple locations delivery. In order to assess the concept of multiple delivery locations, both models are compared with each other.

A Healthy Diet: Different Perspectives

Current diet models are commonly formulated from consumer perspective. However, what might be an optimal healthy diet for consumers might be far from optimal for supermarkets. In this research, it is assumed that the objective of consumers is to minimize the cost of their diet and that the objective of supermarkets is to maximize their total margins. The difference between consumers and supermarkets is modelled by using these two objective functions on three different price levels: 1. Premium brands, 2. Private label, 3. Cheaper alternative. An experiment shows that the chosen objective function has an impact on the intake of individual products, but the total intake of a product group hardly differs between the diets. So, there is substitution between products within the same product group in order to find the optimal solution. This research shows how diet modelling can be used to show differences between diets when one looks from different perspectives.

Reducing food waste in supply chains by intelligent packaging

The aim of the thesis was to estimate the impact of intelligent packaging (IP) on reducing food waste in supply chains. In IP, materials monitor the condition of packaged food or the surrounding environment, and provide information on the freshness of the food.The research focussed on fish, in particular fresh cod fillets sold by retailers in the Netherlands. Literature research showed various opportunities for waste reduction by using the information provided by IP. The most promising opportunities are the use of a dynamic expiry date (DED), dynamic pricing and the adjustment of the distribution according to Quality Controlled Logistics (QCL) from the moment of packaging onwards in the chain. The consumer behaviour was also expected to change positively in case IP is implemented. The current situation without IP was resembled for a simplified supply chain model. Simulation of the effect of IP and using DEDs resulted in a waste percentage of 7%. Adjusting the distribution according to QCL in combination with DEDs is promising. However, many assumptions were necessary at this stage. It is recommended that further research focusses on creating a general optimization model for minimizing waste by making adjustments in the distribution using IP, thereby eliminating the need for assumptions on crucial values.

Synchromodal Freight Transport - a review on main enablers

The basis of this study was to identify and categorize the main enablers to transit from an intermodal to a synchromodal transport system. Since the transport sector is becoming a complex system with excessive demand and causing negative externalities, it is essential to facilitate on creating a flexible and sustainable network, to optimize transport processes whilst using all available resources and multiple modes. Based on that, this research underwent through a literature review on the main enablers, a realization of interviews and a questionnaire development, in order to identify all the main enablers for this system. The different enablers were compiled, analysed and summarized in different categories. The categories for the final enablers evolve technical, regulatory, organizational, cultural and operational elements. In this paper, it turns evident the importance to know in which factors and what are the enablers to give attention when implementing and improving a synchromodal transport system, since only this way, the logistics chain can be operated optimally with the service schedules.

Getting started with the optimization suite Xpress

Xpress is a professional optimization suite for modelling and solving mathematical programming problems. The problem is that the learning curve for using the software takes quite an effort and requires a lot of time for unexperienced users. After following basic courses in programming languages and some courses in decision science, I decided to combine the knowledge of both fields to bridge the gap between an overflow of advanced technical information, scattered around in several professional manuals and a first manual on an entrance level, containing basic information for getting started with the optimization suite Xpress. A well-considered set of basic topics are formulated in advance. All topics are addressed and illustrated in Xpress from small-scale examples in a “Getting started manual”. The goal of the manual is to construct a paper which upcoming students can use to program their thesis work in Xpress.

Optimisation of the safety factor for a perishable product

When considering perishable products, the trade-off between waste and shortage is important. A safety stock functions as a buffer for uncertainty in demand during a week, and will influence the amount of shortage and waste. The purpose of this study was to conduct research into the consequences of using a certain rule of thumb related to order policies for supermarkets facing non-stationary demand of a perishable product.

In this research, three rules of thumb for setting the order-up-to-level  were tested and evaluated by simulation in Excel. The first rule, 1S1X, consists of a constant order-up-to-level that is not weekday dependent with one safety factor, as if demand is stationary. The second rule, 6S1X, considers six different order-up-to-levels for the different weekdays, optimised by one safety factor. The most extensive rule, 6S6X, also considers six order-up-to-levels, now optimised with weekday dependent safety factors.

The research shows that when the 6S6X rule is used for replenishment when facing non-stationary demand, the costs of shortage and waste were lowest for most experimental conditions. Surprisingly, having six order-up-to-levels optimised by one safety factor (6S1X-rule) is worse than a constant order-up-to-level.

The EOQ model for perishable products

This BSc thesis is a follow up on the MSc of Sijtsma (2016). In the thesis of Sijtsma (2016) a variant for the Economic Order Quantity model (EOQ) was developed for mangoes by adding the storage temperature to the model. The aim of this thesis is to find out if there has been done research about adding product decay to the EOQ model other than the research of Sijtsma (2016) and to find out what data should be collected to confirm her adapted model with real data.

In the first part, some of the articles from the literature overview from Bakker, Riezebos and Teunter (2012) combined with new found articles are transformed in a new literature overview. This desk research shows that there are a lot of general extensions to the EOQ model to make it suitable for deteriorating products. However, knowledge about EOQ models for specific products or with the application of microbial and enzymatic degradation is scarce.

In the second part of the research, the extended model of Sijtsma (2012) is explained, followed by a part where the required information is presented and illustrated from a small scale example. Before getting the information it’s important to know how the transportation costs are calculated and how many supplier there are. The fitted constants are not adaptable to the situation yet, since it is unknown how to calculate them. This, together with the estimated data and little transportation time, are the reasons why the results of the calculations are a bit high.

Further research is needed to develop more EOQ models which implement microbial and enzymatic degradation and are modelled for specific product(categorie)s. It is also recommended to do some calculations with the adapted EOQ model of Sijtsma (2012) with data from real life cases.

Optimization of inulin fractionation by modelling membrane cascades - Heuristic approach for solving the pooling problem

An extensive study of the optimization developed by Braakman [10] for a 3-stage membrane cascade for inulin fractionation and the solution methods for this problem is presented in this thesis. This model has been upgraded to include the pooling by problem by considering recycle streams that are mixed before stage 1. As a generalized pooling problem, it may be formulated as a bilinear program (BLP). The validation of this model raises some conclusions: the optimization model accurately predicts the behavior of the system, being the average deviation between the experimental data and the optimization model around 10% for the flow rate, and 6% for concentration in retentate stream 3 (R3). However, some pitfalls have been identified that can be further analyzed to determine the reasons for the variability observed. Finally, a review of the solution methods that have been proposed for the pooling problem, is described in this thesis. The most promising method to applied would be an MP-based heuristic (ALT), that would be relatively easy to implement without requiring too much understanding and yielding reasonable solutions.

Implementation of omni-channel retailing in grocery stores - A literature review

The grocery sector is lagging behing in omni-channel retailing. The aim of this thesis is to identify existing gaps in literature concerning omni-channel grocery retailing and decide if omni-channel retailing easily can be applied in grocery stores. In omni-channel retailing channels are integrated with each other to provide a seamless world of shopping experience. Data for this study is obtained by conducting a literature review. The results show that there are gaps in revenue models, environmental impact and customer satisfaction. Omni-channel grocery retailing is not an easy implementation, otherwise it would be integrated more already, but at the moment there is too little research done. This thesis identified research gaps that need to be investigated for omni-channel grocery retailing to be implemented more easily.

Optimizing crop allocation in Dutch agriculture - Optimizing land-use and crop allocation accounting for price and yield uncertainty

This BSc thesis is about land-use optimization in Dutch agriculture. Optimizing the use of resources is nowadays very important, the world’s population is growing and the area of agricultural ground is decreasing. Farmers are more exposed to the world market and their prices, quota on products like sugar beets and milk have been abandoned. This puts our farmers in a somewhat unreliable position. Therefore the aim of this thesis was to see if it is possible to make a model taking a price and yield uncertainty into account.


In this BSc thesis this problem is tackled using a linear programming with some stochastic variables. First a small literature review has been done to see what has already been done on this topic and to determine a proper methodology. Following this literature review a the model is made. The results following retrieved from the model are analysed and discussed.

Multi-criteria analysis of the adaptation of reverse logistics to an expanded deposit-refund system in the Netherlands

The deposit-refund system (DRS) in the Netherlands already provides the possibility to collect and recycle waste products such as large PET-bottles, glass bottles and beer crates. An expansion of the system to also include small PET-bottles and cans, would have several benefits to the economy, environment, and society. An expansion of the DRS would require retailers to adapt and reorganize their reverse supply chains. This thesis aims to provide an overview of possible new systems for this and of characteristics of retailers that should be considered in the reorganization of reverse supply chains. Two opposing systems are conducted and evaluated on the basis of several criteria using data from existing literature. By matching the characteristics of retailers with these systems, an insight is gained into how retailers can effectively adapt to a new DRS in the Netherlands. The results are applied to a practical example of a retailer. This application, in combination with an interview with the retailer, shows how the results of this thesis can be used by retailers in the industry.

Improving commonly applied deterministic yield models in the arable farming domain using stochastic programming and big data

In this BSC-thesis the added value of a stochastic decision support model is compared to a deterministic linear decision support model to investigate if there is added value in using a big data supported stochastic decision support model in corn farming. To illustrate the added value of using (big) data in a stochastic decision support model a dataset of the NASS is used to derive yield scenario’s and yield probabilities. This resulted in a static yield probability of either bad (chance p = 0.30), normal (chance p = 0.50) or good (chance p = 0.20). The yield is determined by the mean and a deviation of 20% above and below the mean. The outcome of running both models is captured in a profit table and by looking at the generated solutions for first stage, second stage variables and the expected profit a comparison is made. These results are visualised in a graph to allow a clear overview. Based on the profit table it can be concluded that using (big) data in a supported stochastic decision support model results in a better performance compared to the deterministic linear model based on total expected profit. This approach illustrated the added value of using (big) data to reduce uncertainty by determining yield scenario’s and probability. It can be concluded that when available, big data can be used to improve predictive yield models, predictive feed intake models and real-time operational decisions.

The development of the French fries case for the course Food Production Chains, focus on part D (ORL)

A literature study was performed on the topics of part D of the course Food Production Chains at Wageningen University and Research. These topics all revolve around the field of Operation Research and Logistics and are based on linear programming problems. The topics that are addressed during the course are explicit and implicit programming, minimax(maximin)-programming, if-then restrictions, range constraints, semi-continuous variables, discreet valued variables and either-or restrictions.
Based on the literature study and the acquired knowledge, a case is developed for the students of the course Food Production Chains. This case will be the third in line, so that every year a new case can be presented to the students.
The case is about the production process in the factory of Aviko. On the basis of their real life activities a couple of problems are described and students should solve these problems with the help of linear programming solutions that they learn about during the course. These problems are mainly focussed on sourcing of the raw material (potato) and product selection.
End product of this research is part D of the case for the course Food Production Chains

Optimizing the Ethiopian diet ; Linear programming

The current situation in Ethiopia is critical, as almost half of the children are underfed. To make the problems even worse, overnutrition is becoming an emerging problem as well, together with a lack of important nutrients in the diets.

This study uses linear programming to create a practical, nutritious and healthy diet for Ethiopian women of reproductive age. An Extended Goal Programming model has been created in FICO Xpress IVE to provide practical and pragmatic insights to the Ethiopian population.

To combat the existing and emerging problems, the linear programming model creates a diet which complies to all nutritional requirements. To keep the results practical, the Extended Goal Programming model minimized the differences between the actual consumption and the suggested diet by the model. This forces the solution to be closer to the current habits of the target group. The solution provided a balanced, nutritious and most importantly, a healthy diet, close to the real consumption of both Ethiopians living in rural and urban areas.

Key Performance Indicators of animal welfare in veal chains in the Netherlands

The interest of consumer for animal welfare has been increased in recent years. Urbanization, media, influence of civil society organizations and increase in society’s education and economic level are the reason for that animal welfare has become more important (Koknaroglu & Akunal, 2013). Consumers started questioning under which conditions animals are raised, transported and slaughtered (Koknaroglu & Akunal, 2013). Besides that, they started willing to pay more for products that are animal. This will change the animal production practices in the future. Since the current veal chain in the Netherlands has not been designed to account for animal welfare, animal scientists have to adapt themselves for changing animal welfare rules and regulation. The purpose of this study was to identify key performance indicators when improving animal welfare in production chains. Animal welfare is in this study defined as: living or adaptation of animals without suffering to the environment provided by man (Carpenter,1980). The key performance indicators are cortisol, glycogen, creatine kinase, heart rate, respiratory rate, body weight, temperament, bruises, pH and temperature. These indicators are not quantifiable and therefore the correlation has been searched with variables that can be implemented in a quantifiable model. The veal production chain needs to be evaluated quantitatively in order to make trade-offs when looking for optimizations within the chain. These quantifiable variables are: stock density, travel distance and amount of times the animals have to (un)load during transport.

Reducing food waste within airline catering industry, using a simulation model

Food waste is a global problem, as it concerns food that is removed from the food supply chain to be recovered or disposed. This research contributes to the food waste literature within the inflight catering by focussing on the supply chain of the meals that are offered on board. An efficient distribution system is needed to reduce the food waste that occurs within the inflight catering. Within inflight catering, the problem lies in the fact that many meals remain unused after a flight. A simulation model was developed to analyse the opportunities for the unused meals. This simulation model will improve the current situation; in which the meals are destroyed. In addition, there is a focus on more efficient planning, while also looking at the financial area. This research provides recommendations for preventing the food waste after creating an improved situation, in which less waste of food or less money is lost.

Quantifying the bullwhip effect in a three-stage supply chain model

This research quantifies the bullwhip effect in a three-stage supply chain. The goal of the research is to determine how lead times are influencing the bullwhip effect. First, a literature review is conducted in order to find the causes of the bullwhip effect and to determine how the bullwhip is measured. The literature review showed that the main causes of the bullwhip effect are: order batching, demand signal processing, gaming behavior and price variations. After the literature review, a 3-stage supply chain model is built in which the bullwhip effect is quantified on the basis of lead times.

The model contains a manufacturer, a distributor and a retailer. It is a single item model. The results show that the bullwhip effect is influenced by lead times. As longer lead times require more safety stock. A higher safety stock means a higher inventory, and this allows more fluctuation. Which is reflected in the bullwhip effect.

Assortment planning: optimizing the assortment of supermarkets using a mixed-integer linear programming model

Selecting the right assortment for a supermarket can be very crucial. The aim of this bachelor thesis is show how an assortment optimization model can support the decision-making process of choosing the optimal assortment for a supermarket. Literature study is done to get insight in what type of factors could be taken into account by assortment optimization models. With the knowledge obtained from literature study, factors are selected and a mixed-integer linear programming (MILP) is designed that determines the assortment with the highest profit. With the use of a designed dataset of 200 coffee products the model is applied and the optimal assortment for this product category is calculated by Fico Xpress. The results found from this analysis are that the products with the highest profit margin, highest demand and/or products that need a small shelf space are included in the optimal assortment. This research also found that the profit margin and the demand have the most impact on determining the optimal assortment. Using assortment optimization models are very effective in finding the assortment with the maximal profit and there seems to be a clear relationship between the profit and the amount of products in the assortment.