Deformable Image Registration (DIR) is a process in which the point-to-point correspondence between two or more medical images is estimated. This could allow spatial data to be transferred between these images, easing the work of practitioners in the field of radiation oncology. Many DIR approaches already exist. These are not yet applicable in all clinical settings, however, as DIR is a complex problem, partly because of the difficulty of evaluating image registration results. This is why often multiple objectives are considered in this evaluation. In this thesis, we combine two separate multi-objective DIR approaches into hybrids while evaluating their performances, with the aim to improve upon the individual approaches, thus reducing the gap toward clinical practice. In these hybrids, we also use an approach called Automated Landmark Detection which automatically detects corresponding landmarks between the source and target images.

The first DIR approach, the Digital Phantom, creates a biomechanical model of the patient and uses an Evolutionary Algorithm (EA) to optimize parameters of a Finite Element Method (FEM)-based simulation. The second DIR approach, MOREA, creates a dual-dynamic biomechanical transformation model of the patient, using an EA to optimize its parameters.

In the first hybrid approach, called the Landmark Guided Digital Phantom (LGDP), the Automated Landmark Detection approach is used to automatically find corresponding landmarks, which are then used as a guidance objective in the Digital Phantom. The second hybrid is called DP-MOREA, where the output of the Digital Phantom is used to initialize deformations in the MOREA approach. The final hybrid is called LGDP-MOREA. This hybrid combines all three components, using the generated landmarks to guide the Digital Phantom and initializing MOREA with the result. We analyze the effect of these hybrids on the quality of the final registration, both quantitatively and qualitatively. Additionally, we compare these hybrids with the components they are composed of.

We test the performance of the approaches and hybrids on pelvic CT scans of three cervical cancer patients with large deformations. We observe that the individual Digital Phantom component outperforms the LGDP hybrid both quantitatively and qualitatively on these three patients. In addition, we observe that initializing MOREA with the output of the Digital Phantom and LGDP improves on how fast the optimization converges, that it improved the end results quantitatively and that it also has an effect on the resulting Deformation Vector Fields (DVFs). We conclude that in future research, these effects on the DVFs and their clinical relevance should be studied in more detail.

The continual increase in cyber crime revolving blockchain applications calls for secure blockchain systems and clarity on the division of security responsibilities. This research is an integrated project between two master programmes at the Delft University of Technology: Computer Science and Communication Design for Innovation, and focuses on software testing and security responsibilities.

In this study, we investigate if grammar-based fuzzing, a popular approach for identifying bugs in software, is effective on JSON-RPC systems like blockchain applications Ripple and Ethereum. Furthermore, we evaluate whether we can improve upon traditional grammar-based fuzzing by using evolutionary search.
We introduce GEFRA, a black-box grammar-based fuzzing tool that generates tests for JSON-RPC APIs.
Using a diversity-based fitness function that leverages system feedback, GEFRA is able to effectively guide the search process towards new test cases that obtain additional test coverage.

Additionally, various perspectives on blockchain security responsibilities are investigated. A media content analysis was performed and interviews were conducted with legal and blockchain experts.
News media frequently frame end users as responsible for the prevention of blockchain attacks. While attackers are legally responsible, users are left to deal with the consequences if attackers cannot be found. Responsibilities generally end up with users as decentralisation is the core idea of blockchain. Legislation may be the only solution to define a clear division of responsibilities.

In recent years, recommender systems have become a fundamental part of our online experience. Users rely on such systems in situations with many potential choices, such as watching a movie on a streaming service, reading a blog post, or listening to a song. Traditionally, these systems use techniques such as collaborative filtering and content-based recommendation. Both approaches have disadvantages, so to reduce those, recent research combines various techniques in different ways to create hybrid recommender systems. Creating a well-performing hybrid recommender system generally requires extensive knowledge of recommender systems, the domain on which one wants to provide recommendations, and trial and error. Automating this process makes recommender systems accessible for organizations that lack the resources to build these systems themselves. However, there is a lack of research regarding automating this process. This study aims to provide an initial exploration into this area by proposing RACE:GP, an end-to-end approach that automatically produces accurate non-trivial hybrid recommender systems with only a dataset and a definition of 'relevance'. RACE:GP automatically creates a programming language from a dataset in which any valid program is a recommender system on that dataset. By defining the relevant interaction, it can automatically evaluate the accuracy of these programs. It uses a search strategy based on genetic programming to find the best performing recommender systems in the language. To test our hypothesis, RACE:GP is used to produce recommender systems on three well-known datasets in recommender systems literature, and the results are compared to baselines based on collaborative filtering. Additionally, to verify its adaptability, we analyzed the produced recommender systems given different recommendation scenarios. The results showed that RACE:GP is able to produce recommender systems that outperform our chosen baselines by a significant margin. Furthermore, analysis of the produced recommender systems on different recommendation scenarios within a dataset shows that it can find systems that are especially accurate in situations with different densities or recommending specific interactions in datasets that contain different interactions. These results suggest that RACE:GP is a viable and generally applicable approach that makes creating hybrid recommender systems accessible for anyone with a dataset.

This thesis investigates the optimization algorithm used for steel temperature control on the run-out table at Tata Steel IJmuiden. The system currently implemented in the finishing mill is called STORM (Smart Temperature Optimization on the Run out table for Mechanical property control) and was created and fully implemented this year. This control system computes one setup based on the estimated maximum speed and properties of the head of the slab at the time of entering the finishing mill and is based on a Genetic Algorithm (GA). To accommodate for variations over the length of a steel slab, manual action is required. In order to automate this process, the aim is to further develop the system to enable computation of setups for multiple samples over the length of the slab within the limited time available between a slab entering the finishing mill and the start of the cooling process. For this purpose, this thesis introduces three precomputation extensions that incorporate the concept of a warm start with different methods to incorporate knowledge of the problem: STORM-Trained on a Single Model, STORM-Trained on Multiple Models and STORM-Trained on Multiple Objectives. The baseline is taken to be STORM-Uninformed, that uses no prior knowledge to compute multiple setups. The new methods are compared based on four different product recipes with respect to the fitness of the optimal setup found by the solver and the number of iterations required to obtain a useful setup. A setup is assumed to be useful when it is at most 1.2 times the known optimum of a problem instance. The experiments conducted to compare the three mentioned extensions show that when optimizing for samples over the length of the run-out table STORM-TMM outperforms both other methods, though for problem instances with two domains a population size of 200 is required for precomputation and even then not always a useful setup is found within the iteration limit. For optimization for different sets of objectives there is an even greater distinction in the results for problem instances with one and two domains. For the former, the multi-objective algorithm STORM-TMO most definitely performs best, where for the latter STORM-TSM performs best. Again when computing a setup for two domains, a useful setup is not always found. In conclusion, the multi-objective methods STORM-TMM and STORM-TMO can be used for sample setup computation for problems with one domain after some final checks. For problems with multiple domains, additional research is required.

The rail network in the Netherlands is one of the busiest in Europe. To ensure a safe and reliable infrastructure, preventive maintenance is of utmost importance. ProRail, the sole maintainer of the railway infrastructure in the Netherlands, spends hundreds of millions of dollars on maintenance each year. Due to the complexity and busyness of the network, maintenance can cause major disruptions leading to longer travel times. Despite these factors, maintenance is currently scheduled mostly manually, leading to suboptimal schedules being created. A pilot study by Macomi, the company which enabled this study through an internship, showed great potential for improvement in scheduling this maintenance. In this thesis, the aim is to further improve the maintenance schedule of ProRail, as well as to find out what kind of solution method is most suitable to improve that schedule. To achieve this goal, various types of algorithms have been implemented and tested on the problem. An evolution strategy algorithm developed by Macomi was successfully improved. Through the use of multi-objective algorithms, the trade-off between maintenance costs and availability of the infrastructure was analyzed. These multi-objective algorithms were found to provide unsatisfactory solutions. Greedy algorithms were also developed to provide a quicker solution method, and the resulting solutions were of surprisingly high quality. Finally, a hybrid algorithm using the evolution strategy and a greedy algorithm was created. It was shown that this hybrid algorithm outperforms all other algorithms, and provides solutions that are better in terms of costs and constraints compared to the manual schedule of ProRail, and the baseline established by the pilot study of Macomi.

In many real world scheduling problems there exist hard deadlines after which tasks can no longer be performed. Conversely, not all tasks are necessarily required to be scheduled. Furthermore, the problem investigated in this thesis includes sequence dependent setup times, an aspect reminiscent of the Travelling Salesperson problem. These elements are the underlying additions of the Order Acceptance and Scheduling with Sequence Dependent Setup Times, and introduce additional difficulties to the problem of scheduling. Given the complexity of this problem, it is surprising no evaluation of black-box approaches has been performed. This thesis investigates the applicability, strengths and weaknesses of black-box optimization approaches to various instances of this problem. Two hints were provided in order to improve performance initially. The Time hint informs the approaches of the completion time of the last order in the evaluated schedule, while the Bounds hint leaks information about the release time and deadlines of each order. Both have notable up- and downsides with regards to performance. For the identified weaknesses improvements to the most promising approach, Permutation GOMEA, are proposed and evaluated. These improvements include the addition of various local search approaches, including one derived from an exact approach. The most notable result is a new approach based of GOMEA, named qGOMEA, which shows performance beyond that of any of the evaluated algorithms, including the current state-of-the-art of the Order Acceptance and Scheduling problem. Furthermore, applicability to other permutation problems was shown by improved performance on benchmark functions.

District heating systems (DHS) are considered the best alternative to individual heat boilers as they have higher efficiency and enable the use of sustainable heat sources, like geothermal heat sources or waste heat from industry. Currently, most DHSs are operated by choosing a temperature once every day depending on the weather. This setting is chosen such that peak demand can be satisfied. However, when demand is lower, the supply temperature will be higher than necessary and heat will be wasted. In addition to this, heat production costs can be dynamic over time, which allows more cost-efficient heat production scheduling. By choosing a dynamic temperature over the day losses and operating costs can be reduced. However, determining these dynamic temperatures is not easy, as there are several factors that need to be taken into account to ensure that demand can always be fulfilled. In this research, the use of metaheuristics is explored for finding supply temperatures. Optimisation of DHS operations results in up to 5% savings with respect to typical heating curve operations. The second contribution is a method to determine a theoretical lower bound on the operating costs of a DHS, as this was not yet found in literature.