05 Fakultät Informatik, Elektrotechnik und Informationstechnik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/6
Browse
12 results
Search Results
Item Open Access Flexible and efficient data mapping for simulation of coupled problems(2026) Schneider, David; Uekermann, Benjamin (Jun.-Prof. Dr.)Multi-physics simulations model various physical phenomena and their interactions. Examples include climate models or the simulation of fusion reactors. Modeling more physical phenomena in the same simulation often provides new insights. This poses significant challenges for the underlying methods and the simulation software itself. Decomposing a multi-physics simulation into its parts is an effective way to tame the inherent complexity. The coupling library preCICE allows for such partitioned simulations, coupling simulation models without access to their internal numerics. While preCICE is well-suited for conventional, mesh-based surface couplings, its applicability breaks down in alternative scenarios due to restrictive data-mapping algorithms. These algorithms apply spatial mapping operators to statically defined coupling meshes and are generally not designed for large problems. As a result, they hinder the flexible and efficient simulation of relevant applications, including volumetric couplings, high-order couplings, and mesh-particle couplings. To overcome these limitations, this work focuses on four main aspects: first, developing a scalable partition-of-unity radial-basis-function interpolation customized for coupled problems; second, implementing data-parallel kernel methods on CPUs and GPUs to ensure cross-platform efficiency; third, enabling immediate access to meshes received from coupling partners for user-defined mapping operators; fourth, computing a mapping operator just-in-time on temporary coordinates for the seamless coupling of meshless solvers. Taken together, these concepts enable plug-and-play integration of diverse numerical models in multi-physics simulations. Large-scale volumetric couplings are now feasible and efficient, breaking the traditional accuracy-efficiency trade-off. Multi-physics couplings can exploit spatial high-order convergence rates of existing models for high-fidelity simulations, while preserving full black-box compatibility. The simulation of fluid-particle couplings can be modularized and leverage already-existing models for both the simulation of the mesh-based fluid and the particles. In the end, the robust implementation, the gained efficiency, and the flexibility significantly extend the applicability of preCICE and benefit its vibrant user community. Beyond preCICE, the presented concepts provide generally applicable building blocks for scalable, modular multi-physics coupling.Item Open Access Data engineering concepts, framework, and algorithms for personalized and transparent decision support(2025) Oppold, Sarah; Herschel, Melanie (Prof. Dr.)Machine learning models are commonly used for decision support even though they are far from perfect, e.g., due to bias introduced by imperfect training data or wrong feature selection. While efforts are made and should continue to be put into developing better models, we will likely continue to rely on imperfect models in many applications. In this thesis, we follow a rationale similar to the practices that emerged in medicine to develop decision support systems (DSS) that are demonstrably developed as responsibly as possible. Therefore, we present novel approaches to develop personalized and transparent data-driven decision support systems. Our fist contribution is a novel holistic system framework that offers transparent and personalized services tailored to user profiles to serve their best interest. Our framework personalizes the choice of model for individuals or groups of users based on metadata about datasets and machine learning models. Querying and processing these metadata ensures transparency by supporting various kinds of queries by different stakeholders. We discuss our framework in detail, show why existing solutions are inadequate, and highlight research questions that need to be addressed in the future. Based on a prototypical implementation, we showcase that even a baseline implementation of our framework supports the desired transparency and personalization. To address the problem of personalization, we present novel algorithms for dynamic, fair, and accurate decision support systems in our second contribution. We first propose a general procedure that combines context-sensitive algorithms for generating dynamic model ensembles necessary for personalization with (static) fair and accurate model ensembles, satisfying our multi-objective goal. We further introduce a family of algorithms, jointly called FALCES, representing different alternatives for our procedure. Using the FALCES algorithms, we evaluate our framework for dynamic, fair, and accurate model ensembles on synthetic and real data. The results show that despite the presence of biases, our algorithms outperform state-of-the-art fairness algorithms while maintaining acceptable accuracy. In order to provide personalized and transparent decision support, we have to capture and manage a large and diverse set of metadata on datasets, which are the foundation of decision support systems. Therefore, we present LiQuID in our third contribution. LiQuID is the first systematic and holistic metadata model for accountable datasets, i.e., datasets on which queries of an accountability workload can be answered. We evaluate LiQuID by comparing it to existing metadata standards from the responsible data analysis community and more matured disciplines. In a second evaluation, we compare LiQuID to a new workload for accountable datasets that we created based on a vast survey of the GDPR, an FTC report, and dataset expert interviews. Finally, we look at the devised system’s usefulness in achieving the desired goals. In our fourth contribution, we delve into one particular setting that has been claimed by research before, where transparency potentially serves to improve trust. We critically review the term “trust” to define a theoretical model for trust in data engineering. Based on this model, we describe a framework for trust in data engineering that integrates trust in the data engineering pipeline and serves as a guideline to develop a trust strategy. We also describe a general procedure to evaluate the effects of a transparency measure on trust. Finally, we apply and evaluate our methods on a credit scoring use case. Results show that transparency not necessarily increases trust, highlighting the importance of a more systematic study of the problem using our proposed methods. The research presented in this thesis has significantly contributed to our goal for decision support systems that are demonstrably developed as responsible as possible. This includes a general system framework for personalized and transparent decision support, novel optimization algorithms, accountability metadata models and a critical discussion on using transparency to foster trust.Item Open Access Dynamic quality of service concepts for wired and wireless networks(2025) Laidig, Robin; Rothermel, Kurt (Prof. Dr. rer. nat. Dr. h. c.)The unprecedented increase of networked devices through the Internet of Things paved the way for Networked Control Systems (NCS), which are found all around us today. NCSs control physical processes over computer networks, for instance, in smart homes, smart factories, automobiles, or drone operations. While it is cost-efficient and flexible to use computer networks, the often stringent latency requirements of NCS make Quality of Service (QoS) models for the network communication essential. QoS models provision network resources and can be used to ensure latency and jitter guarantees for time-sensitive real-time traffic, or to improve reliability and fairness for best-effort network traffic. In order to provide latency guarantees, current QoS models require a strict traffic specification from the application. They then distinguish between two classes of network traffic: compliant traffic, which conforms to the traffic specification and receives the latency guarantees, and non-compliant traffic, which receives no guarantees at all. However, it is difficult to accurately predict and specify the network traffic of applications, especially for NCS that interact with the physical world. Moreover, the static nature of traditional QoS models, which only distinguish between compliant and non-compliant traffic, makes it dangerous to underestimate an application's traffic specification, as non-compliant traffic will lose its QoS guarantees, which could have catastrophic consequences for the safety of the system. Therefore, there is currently a gap for dynamic QoS models, that can react to and mitigate unexpected network conditions. In particular, they should be able to compensate for short-term traffic that is not conforming to the traffic specification (out-of-spec traffic) without losing all QoS guarantees. Furthermore, dynamic QoS models that can adapt to fluctuating network conditions are required for NCS with wireless networks, such as drone operations. In this thesis, we address the problem of unexpected network conditions that can occur when working with NCS. To this end, we investigate dynamic QoS models for both wired, time-sensitive networks and wireless, best-effort networks. In particular, we first address the problem of compensating for short-term non-compliant traffic in wired, time-sensitive networks, such as that of NCS. We propose the Dynamic Deterministic QoS model, which gracefully degrades QoS guarantees in case of short-term violations. It is implemented by the Dynamic Priority Token Bucket (DPTB), which is a proposed token bucket extension designed to work on top of existing scheduling algorithms, such as the IEEE Asynchronous Traffic Shaper. DPTB does not require any modifications to devices inside the network, only at the edge, and can provide weaker, yet still deterministic latency guarantees for excess packets. These provided sub-guarantees improve control performance and resource utilization efficiency of NCSs, which we demonstrate using a proposed physical NCS benchmark setup. Second, we address out-of-spec traffic in the form of delayed frames in time-based networks. We propose a frame elevation policy for the time-slotted IEEE Time-Aware Shaper (TAS) to enhance it with weakly-hard real-time capabilities. %The TAS is a popular QoS mechanism for providing real-time latency guarantees over Ethernet. By selectively elevating the priority of some delayed frames instead of dropping them, they are forwarded with highest priority to catch up with their original timeslot and meet their deadline. Implemented by a post-processing routine for existing TAS schedules, we propose modifications such that the original deadlines still hold even when elevated frames interrupt scheduled transmissions. We show that our elevation policy can be used to provide weakly-hard real-time guarantees for NCSs, with combined wireless and wired networks, using our physical benchmark setup. Third, we address the problem of fluctuating network conditions, caused by obstacles or frequency jamming attacks in wireless networks used for drone communication. A drone operation is an NCS, where the drone receives control commands from the ground control station (GCS) and replies with status updates over a wireless network. However, wireless communication differs from wired communication in that it is often error-prone and subject to interference or even jamming attacks. Therefore, we develop an adaptive drone communication redundancy (ADCR) mechanism that detects communication outages and mitigates them by dynamically adapting communication over heterogeneous, redundant network interfaces. We show that ADCR balances the trade-off between reliability and efficiency and provides similar reliability as fixed redundancy but at a significantly lower overhead. Fourth, we address fluctuating network conditions in the form of excessive latency in large-scale, public WiFi networks. Such public WiFi networks, for example, found in airports or public transportation today, often provide unsatisfactory QoS, caused by congestion delays. To this end, we propose a usage-dependent QoS model (UD-QoS) that dynamically classifies the network behavior of wireless devices and prioritizes those that cause lower network loads in case of congestion. We show that this shields interactive devices with low network usage from high latency, caused by congestion of packets from devices with high network usage. In addition, this dynamic QoS model provides an incentive for users to self-regulate, in order to receive better QoS in congested networks.Item Open Access Scalable traffic engineering heuristics for time-triggered communication in real-time networks(2026) Geppert, Heiko; Rothermel, Kurt (Prof. Dr. rer. nat. Dr. h. c.)Distributed safety-critical cyber-physical systems require real-time behavior. This means they must respond not just quickly, but in time, to new situations considering both, the task processing and network communication time. From a networking perspective, meticulous, time-driven traffic planning performed at the frame level is necessary to guarantee low end-to-end delay bounds and low latency. This involves carefully planning transmission operations along each time-critical frame's network path are carefully planned, including precise timing, to limit or even eliminate interference from cross-traffic and ensure timely delivery. Since modern real-time systems can consist of hundreds or thousands of devices - for example, large manufacturing plants or continental-sized power grids - the traffic planning must be highly scalable. Although there are many traffic planning approaches in the literature, there is a lack of very fast heuristics that can handle very large stream sets and networks quickly. This thesis investigates traffic planning heuristics and optimization techniques, focusing on different aspects of the traffic planning domain. The traffic planning consists of novel methods for conflict-graph-based scheduling and new heuristics for very large instances of traffic planning problem. The optimizations include multicast partitioning, which combines the benefits of multicast and unicast traffic plans, and load-balanced stream placement, which generates traffic plans that can accommodate additional streams joining the system later. We created prototype implementations and analyzed their performance in solving the traffic planning problem. Our traffic plans yielded a higher accumulated network throughput or admitted more streams while maintaining computation times ranging from sub-seconds to minutes, even for extremely large-scale problem instances. The traffic planning methods and optimization techniques presented in this thesis can be applied to modern real-time networking technologies, such as Time-Sensitive Networking and TTEthernet.Item Open Access From irregular parallelism to portable GPU kernels : enabling efficient task-based GPU programming with HPX, Kokkos and CPPuddle to accelerate stellar mergers(2026) Daiß, Gregor; Pflüger, Dirk (Prof. Dr.)Adaptive, tree-based structures are the foundation of many of the most efficient algorithms and applications. Yet, running such applications on supercomputers is challenging, as the application developers have to handle, for example, distributed tree-traversals and load balancing. Distributed, asynchronous many-task runtime systems such as HPX aim to alleviate these challenges by embracing fine-grained tasks, enabling a fine interweaving of communication and computation through their task-graph. However, the fine-grained approach embraced by HPX is the antithesis for implementing efficient GPU kernels, which leverage the large numbers of available parallel work items to scale to all compute units of a GPU and to hide latencies. This is compounded by the synchronization of GPU operations, which necessitates blocking entire CPU threads, nullifying any potential benefits of interleaving CPU tasks with the GPU operations through the HPX task-graph. Hence, efficiently leveraging GPU-accelerated supercomputers with HPX applications can still be challenging, albeit for different reasons. In this work, we reconcile these differences and unlock the computational performance of GPUs for HPX applications. Our contributions can be used in any HPX application. However, to benchmark them, we turn to a specific real-world HPX application, the astrophysics code Octo-Tiger. Octo-Tiger is used for large-scale simulations of stellar mergers and demonstrates both the pitfalls and the potential of leveraging large GPU-accelerated supercomputers for applications that rely on such a task-based approach. We refactored Octo-Tiger's CPU-only code, ported its computational hotspots to GPUs, and ultimately turned it into a GPU-accelerated application, greatly speeding up these hotspots with various new GPU compute kernels. These new GPU kernels achieve speedups of up to $283$x compared to Octo-Tiger's previous implementation. However, they also demonstrate the pitfalls of simply replacing CPU tasks by GPU kernels, as Octo-Tiger's overall runtime initially increased almost eightfold when using them despite these individual speedups within the hotspots. To solve the limiting problems, we employ various techniques. Most notably, we integrated asynchronous GPU operations into the HPX task-graph using a polling approach, dynamic GPU kernel fusion based on specialized executors, and GPU memory pools that are optimized for HPX. With these solutions in place, we translate the individual speedups of the GPU kernels into tangible speedups for the entire application of up to 8.5x. Furthermore, we combine this approach with the performance portability library Kokkos, to be able to target different GPU-accelerated supercomputers. In this work, we particularly demonstrate scalability on the supercomputers Perlmutter and Frontier, distributing Octo-Tiger across over a thousand GPU compute nodes.Item Open Access Ansätze für flexible und fehlertolerante modellgetriebene IoT-Anwendungen in dynamischen Umgebungen(2024) Del Gaudio, Daniel; Mitschang, Bernhard (Prof. Dr.-Ing. habil.)Item Open Access Surrogate modeling with scientific machine learning(2025) Leiteritz, Raphael; Pflüger, Dirk (Prof. Dr.)Item Open Access Data-driven optimization using deep reinforcement learning with applications in semiconductor testing and digital healthcare(2025) Domanski, Peter; Pflüger, Dirk (Prof. Dr. rer. nat.)The increasing availability of data in areas such as semiconductor testing and digital healthcare, enabled by advances in measurement systems, sensors, and wearables, offers new opportunities for data-driven approaches. However, applications in these domains are often performance-critical and must satisfy strict constraints, such as time consumption, resource usage, and robustness. Addressing such challenging task-specific constraints typically requires optimizations regarding the metrics of interest. Existing optimization methods, which are general-purpose algorithms designed for a wide range of optimization problems, are not tailored to the structural properties of a specific domain. Consequently, existing methods are often computationally costly and time-consuming, making them unsuited for many real-world challenges in performance-critical domains, such as real-time optimization and low-latency inference of diagnostic models in semiconductor testing and digital healthcare. This thesis investigates data-driven, learning-based optimization strategies that are trainable to domain-specific challenges and designed for low time-to-optimize. A key difficulty in many real-world settings is the cost of labeling measurement data with ground truth information. In environments involving systems such as Automatic Test Equipments (ATEs) in semiconductor testing or clinical simulators in digital healthcare, interactive feedback loops can be implemented to enable the training of optimization algorithms directly through environment interactions. Therefore, Reinforcement Learning (RL) is a viable method for training, allowing the use of the feedback metrics directly as a reward signal to tailor optimization algorithms, represented as Machine Learning (ML) models, to the given domain. Once trained, these algorithms provide fast optimization and high performance, making them well-suited for tasks such as Post-Silicon Validation (PSV) tuning and diabetes treatment planning, where frequent optimization of closely related tasks is required. This thesis further explores the use of Large Language Models (LLMs), which have recently demonstrated impressive capabilities in natural language-related tasks such as code generation. Code generation is a relevant task in domains such as semiconductor testing. While transformer-based LLMs have limitations in numerical optimization due to their focus on natural language processing, they exhibit strong capabilities in generating high-level source code, which is particularly valuable in semiconductor testing, e.g., to generate test cases that target specific non-functional properties of a given Device Under Test (DUT). By combining LLMs with RL-based fine-tuning, pre-trained models can be adapted towards task-specific goals such as generating test patterns that effectively address non-functional properties of given hardware to trigger hardware-specific faults. Deploying learned optimization methods in real-world systems introduces additional challenges, particularly when targeting resource-constrained or edge devices, as is typically the case in digital healthcare. A successful deployment strategy must meet strict latency, memory, and energy constraints-particularly critical in edge scenarios, such as on-chip performance optimizations or real-time information on wearable healthcare monitors. To address these challenges, techniques such as (hardware-aware) Neural Architecture Search (NAS), quantization, network compression (e.g., pruning and knowledge distillation), and edge-cloud offloading are of significant importance. They can reduce the computational cost of a learned optimization algorithm on the target hardware while preserving performance, enabling the practical deployment of optimization strategies in resource-constrained, embedded hardware platforms. The proposed methods enable domain-specific, data-driven optimization strategies that balance one-time training costs and long-term gains in repeated, large-scale optimization. This thesis presents a promising direction for advancing performance optimization algorithms in complex, data-rich environments with distinct goals and constraints, such as in semiconductor testing and digital healthcare.Item Open Access Stable and mass-conserving high-dimensional simulations with the sparse grid combination technique for full HPC systems and beyond(2024) Pollinger, Theresa; Pflüger, Dirk (Prof. Dr.)In the light of the ongoing climate crisis, mastering controlled plasma fusion has the potential to be one of the pivotal scientific achievements of the 21st century. To understand the turbulent fields in confined fusion devices, simulation has been and continues to be both an asset and a challenge. The main limiting factor to large-scale high-fidelity predictive simulations lies in the Curse of Dimensionality, which dominates all grid-based discretizations of plasmas based on the Vlasov-Poisson and Vlasov-Maxwell equations. In the full formulation, they result in six-dimensional grids and fine scales that need to be resolved, leading to a potentially untractable number of degrees of freedom. Typical approaches to this problem - coordinate transformations such as gyrokinetics, grid adaptation, restricting oneself to limited resolutions - do not directly address the Curse of Dimensionality, but rather work around it. The sparse grid combination technique, which forms the center of this work, is a multiscale approach that alleviates the curse of dimensionality for time-stepping simulations: Multiple regular grid-based simulations are run and update each other’s information throughout the course of simulation time. The present thesis improves upon the former state-of-the-art of the combination technique in three ways: introducing conservation of mass and numerical stability through the use of better-suited multiscale basis functions, optimizing the code for large-scale HPC systems, and extending the combination technique to the widely-distributed setting. Firstly, this thesis analyzes the often-used hierarchical hat function from the viewpoint of biorthogonal wavelets, which allows to replace the hierarchical hat function by other multiscale functions (such as the mass-conserving CDF wavelets) in a straightforward manner. Numerical studies presented in the thesis show that this not only introduces conservation but also increases accuracy and avoids numerical instabilities - which previously were a major roadblock for large-scale Vlasov simulations with the combination technique. Secondly, the open-source framework DisCoTec was extended to scale the combination technique up to the available memory of entire supercomputing systems. DisCoTec is designed to wrap the combination technique around existing grid-based solvers and draws on the inherent parallelism of the combination technique. Among several other contributions, different communication-avoiding multiscale reduction schemes were developed and implemented into DisCoTec as part of this work. The scalability of the approach is asserted by an extensive set of measurements in this thesis: DisCoTec is shown to scale up to the full system size of four German supercomputers, including the three CPU-based Tier-0/Tier-1 systems. Thirdly, the combination technique was further extended to the widely-distributed setting, where two HPC systems synchronously run a joint simulation. This is enabled by file transfer as well as sophisticated algorithms for assigning the different simulation instances to the systems, two of which were developed as part of this work. By the resulting drastic reductions in the communication volume, tolerable transfer times for combination technique simulations on different HPC systems have been achieved for the first time. These three advances - improved numerical properties, scaling efficiently up to full system sizes, and the possibility to extend the simulation beyond a single system - show the sparse grid combination technique to be a promising approach for future high-fidelity simulations of higher-dimensional problems, such as plasma turbulence.Item Open Access Adaptive human-robot policy blending for shared control teleoperation(2024) Oh, Yoojin; Toussaint, Marc (Prof. Dr. rer. nat.)