arXiv subjects

Alexander Mitsos

Publications and source records attributed to Alexander Mitsos.

At least 19 recordsLinked to original sources

arXiv2026

Systematic Evaluation of TabPFN-TS for Zero-Shot Probabilistic Heat Load Forecasting in District Heating Networks

District heating energy hubs require reliable heat load forecasts for efficient operational scheduling. Conventional forecasting workflows train system-specific models on historical data, which can become burdensome when networks change through new consumers, retrofits, or changing operating regimes. Zero-shot time-series foundation models and in-context forecasting offer a promising alternative: they can adapt at inference time from recent observations rather than by repeated retraining. This study systematically evaluates TabPFN-TS against time-series foundation models and trained machine-learning baselines for probabilistic heat load forecasting in district heating networks. Unlike foundation models pretrained on large collections of real time series, TabPFN-TS relies on synthetic pretraining data, which avoids direct pretraining-test overlap but raises the question of whether the learned prior captures district heating dynamics. We analyze covariate choice, context length, temporal resolution, and prediction horizon on representative operating weeks, validate the selected configuration over a full year, and test transferability on a second network. The results identify hourly 24-hour forecasting with a 12-week rolling context and ambient temperature as a parsimonious high-performing configuration; longer context windows do not improve accuracy. TabPFN-TS remains close to Chronos-2 in deterministic accuracy, reaching CVRMSE values of 13.06% versus 12.48% on the main dataset, and lies within the critical-difference threshold in the daily-rank comparison. Although Chronos-2 achieves the lowest aggregate full-year error, TabPFN-TS shows better empirical calibration. Finally, the diagnostic findings motivate a Multi-Resolution Residual-Correction Forecaster that combines a low-frequency Base Forecaster with a short-horizon Residual Forecaster to improve longer-horizon planning accuracy.

cs.LG↗

arXiv2026

CHAMB-GA: A Containerized HPC Scalable Microservice-Based Framework for Genetic Algorithms

Metaheuristic-based global optimization with embedded, long-running simulations is a computationally expensive process. To support various stages of development and execution, a seamless transition from personal computers to distributed clusters is desired, enabling execution across all computational scales. However, existing tool chains are often characterized by rigidity and hardware-bound constraints, which impede scalability and the integration of complex simulations. Bridging this gap, we present a containerized HPC scalable microservice-based framework for genetic algorithms with embedded simulations (CHAMB-GA). The deployment of the framework scales consistently across cloud infrastructure via container orchestration and HPC clusters via batch-scheduled parallel execution. Users provide the GA operators and simulation backend separately. The framework is designed to run these components in a distributed and decoupled manner, mapped to separate hardware. This approach ensures that the fitness evaluation and genetic operations are not managed within the same process and are utilizing distinct parts of the compute infrastructure. A central message broker coordinates asynchronous manager-worker communication between microservices, thereby parallelizing evolutionary operations and fitness evaluations. We demonstrate CHAMB-GA's scalability, portability, and reproducibility, while facilitating the integration of external tools and complex simulations on benchmark and powerflow problems. The capabilities of CHAMB-GA are validated in a two-part approach: (i) a benchmark study demonstrating minimal overhead while scaling to over 3,500 CPU cores, and (ii) a dispatch optimization of High Voltage Direct Current (HVDC) lines in the German transmission grid, showing seamless migration from Kubernetes to SLURM, combined horizontal and vertical scaling, and integration of multi-stage workflows.

cs.DC↗

arXiv2026

Hybrid Neural Ordinary Differential Equations for Data-Efficient Polymerization Modeling with Incomplete Kinetics

Accurate prediction of polymerization dynamics is essential for process design, control, and optimization. Yet, purely mechanistic models require labor-intensive parameterization of partially characterized kinetics, while purely data-driven models demand large, diverse datasets that are costly to obtain, particularly in early-design stages. We propose a hybrid Neural Ordinary Differential Equation (NODE) framework for data-efficient modeling of free-radical polymerization. Using batch polymerization of methyl methacrylate (MMA) as a case study, the mechanistic mass balances are retained explicitly, and only the partially-characterized effective radical concentration governing monomer consumption is learned from data through a neural network surrogate, while established reactions such as initiator decomposition, propagation, and termination remain physically modeled. The hybrid NODE is evaluated against a discrete-time feedforward neural network and a purely data-driven NODE under sparse data conditions, with models trained on as few as ten measurements under both regular and irregular sampling. The hybrid NODE consistently achieves lower prediction errors and more physically consistent extrapolations than both purely data-driven baselines. In a generalization scenario with noisy data and unseen operating conditions, the hybrid NODE achieves an RMSE of 0.013, compared to 0.31 for the data-driven NODE and 0.68 for the discrete-time model, demonstrating that learning only a closure term rather than the full dynamics is sufficient for reliable prediction under limited data availability.

cs.LG↗

arXiv2026

Robust Design of Multi-Energy Systems Accounting for Mixed-Integer Operational Problems

Identifying robust designs for multi-energy systems is computationally challenging. As rigorous approaches are often computationally intractable, heuristics are employed to generate candidate designs. Specifically, we consider a heuristic that iteratively identifies and adds extreme scenarios to the design problem. We theoretically investigate how three common nonconvexities, i.e., piecewise-linear energy inflow-outflow relationships, minimum part-loads, and storage complementarity, affect the robustness of designs identified by this heuristic. We find that, if surplus energy cannot be curtailed, any of these nonconvexities may cause the heuristic to fail. If curtailment is allowed, storage complementarity does not compromise robustness, and convex piecewise-linear inflow-outflow relationships can be reformulated linearly. However, minimum part-loads may lead to failure of the heuristic. Furthermore, if the optimal value function of the operational problem is nonconvex in the uncertain variables, the heuristic may fail. We demonstrate these findings using an illustrative multi-energy system case study, in which minimum part-loads and nonconvex dependence of the objective function on heat-pump efficiency are identified as possible failure modes. We rigorously verify the robustness of a design identified using the heuristic and find a scenario where, in one time step, 1.1 kW of the 313.6 kW electricity demand could not be satisfied. However, when the number of representative scenarios is increased from 4 to 6 or 8, the resulting designs are robust. This demonstrates that the heuristic can serve as an effective first step in identifying robust designs. However, when robustness guarantees are required, a rigorous solution method must be employed.

math.OC↗

arXiv2026

Iterative Model-Learning Scheme via Gaussian Processes for Nonlinear Model Predictive Control of (Semi-)Batch Processes

Batch processes are inherently transient and typically nonlinear, motivating nonlinear model predictive control (NMPC). However, adopting NMPC is hindered by the cost and unavailability of dynamic models. Thus, we propose to use Gaussian Processes (GP) in a model-learning NMPC scheme (GP-MLMPC) for batch processes. We initialize the GP-MLMPC using data from a single initial trajectory, e.g., from a PI controller. We iteratively apply the NMPC embedded with GPs to run batches and update the GP with new observations from each iteration, thereby achieving batch-wise improvements. Using uncertainty quantification from the GPs, we formulate chance constraints to enforce safe operation to the required confidence levels. We demonstrate our approach in \textit{silico} on a semi-batch polymerization reactor for tracking and economic objectives over durations of two hours, and the reactor temperature is constrained in a range of $\pm2^\circ C$ around its setpoint. After only four batch iterations, tracking error from the GP-MLMPC scheme converged to a reduction of $83\%$, compared to the initial trajectory. Furthermore, under an economic objective, the GP-MLMPC resulted in a 17-fold increase in final product mass by iteration 8, compared to the initial trajectory. In both cases, the resulting GP-MLMPC performance is on par with the full-model NMPC, which shows that the optimal controller can be learned by the approach. By collecting samples around the optimal trajectory, the GP-MLMPC remains sample-efficient across iterations and achieves quick convergence. Thus, the proposed GP-MLMPC scheme presents a promising data-efficient approach for the control of nonlinear batch processes without mechanistic knowledge.

cs.LG↗

arXiv2026

Tabular foundation models for in-context prediction of molecular properties

Accurate molecular property prediction is central to drug discovery, catalysis, and process design, yet real-world applications are often limited by small datasets. Molecular foundation models provide a promising direction by learning transferable molecular representations; however, they typically involve task-specific fine-tuning, require machine learning expertise, and often fail to outperform classical baselines. Tabular foundation models (TFMs) offer a fundamentally different paradigm: they perform predictions through in-context learning, enabling inference without task-specific training. Here, we evaluate TFMs in the low- to medium-data regime across both standardized pharmaceutical benchmarks and chemical engineering datasets. We evaluate both frozen molecular foundation model representations, as well as classical descriptors and fingerprints. Across the benchmarks, the approach shows excellent predictive performance while reducing computational cost, compared to fine-tuning, with these advantages also transferring to practical engineering data settings. In particular, combining TFMs with CheMeleon embeddings yields up to 100\% win rates on 30 MoleculeACE tasks, while compact RDKit2d and Mordred descriptors provide strong descriptor-based alternatives. Molecular representation emerges as a key determinant in TFM performance, with molecular foundation model embeddings and 2D descriptor sets both providing substantial gains over classic molecular fingerprints on many tasks. These results suggest that in-context learning with TFMs provides a highly accurate and cost-efficient alternative for property prediction in practical applications.

cs.LG↗

arXiv2026

Bayesian Optimization of Partially Known Systems using Hybrid Models

Bayesian optimization (BO) has gained attention as an efficient algorithm for black-box optimization of expensive-to-evaluate systems, where the BO algorithm iteratively queries the system and suggests new trials based on a probabilistic model fitted to previous samples. Still, the standard BO loop may require a prohibitively large number of experiments to converge to the optimum, especially for high-dimensional and nonlinear systems. We present a hybrid model-based BO formulation that combines the iterative Bayesian learning of BO with partially known mechanistic physical models. Instead of learning a direct mapping from inputs to the objective, we write all known equations for a physics-based model and infer expressions for variables missing equations using a probabilistic model, in our case, a Gaussian process (GP). The final formulation then includes the GP as a constraint in the hybrid model, thereby allowing other physics-based nonlinear and implicit model constraints. This hybrid model formulation yields a constrained, nonlinear stochastic program, which we discretize using the sample-average approximation. In an in-silico optimization of a single-stage distillation, the hybrid BO model based on mass conservation laws yields significantly better designs than a standard BO loop. Furthermore, the hybrid model converges in as few as one iteration, depending on the initial samples, whereas, the standard BO does not converge within 25 for any of the seeds. Overall, the proposed hybrid BO scheme presents a promising optimization method for partially known systems, leveraging the strengths of both mechanistic modeling and data-driven optimization.

cs.LG↗

arXiv2026

Differentiable Thermodynamic Phase-Equilibria for Machine Learning

Accurate prediction of phase equilibria remains a central challenge in chemical engineering. Physics-consistent machine learning methods that incorporate thermodynamic structure into neural networks have recently shown strong performance for activity-coefficient modeling. However, extending such approaches to equilibrium data arising from an extremum principle, such as liquid-liquid equilibria, remains difficult. Here we present DISCOMAX, a differentiable algorithm for phase-equilibrium calculation that guarantees thermodynamic consistency at both training and inference, only subject to a user-specified discretization. The method combines discrete enumeration of feasible phase states with masked softmax aggregation in the backward pass, with the propagation of the true equilibrium state in the forward pass, using a straight-through gradient estimator to enable physics-consistent end-to-end learning of neural \gls{gE}-models. We show that this approach bears analogy to statistical thermodynamics, and we evaluate it on binary liquid-liquid equilibrium data where it outperforms existing surrogate-based methods, while offering a general framework for learning from different kinds of equilibrium data.

cs.LG↗

arXiv2026

Clapeyron Neural Networks for Single-Species Vapor-Liquid Equilibria

Machine learning (ML) approaches have shown promising results for predicting molecular properties relevant for chemical process design. However, they are often limited by scarce experimental property data and lack thermodynamic consistency. As such, thermodynamics-informed ML, i.e., incorporating thermodynamic relations into the loss function as regularization term for training, has been proposed. We herein transfer the concept of thermodynamics-informed graph neural networks (GNNs) from the Gibbs-Duhem to the Clapeyron equation, predicting several pure component properties in a multi-task manner, namely: vapor pressure, liquid molar volume, vapor molar volume and enthalpy of vaporization. We find improved prediction accuracy of the Clapeyron-GNN compared to the single-task learning setting, and improved approximation of the Clapeyron equation compared to the purely data-driven multi-task learning setting. In fact, we observe the largest improvement in prediction accuracy for the properties with the lowest availability of data, making our model promising for practical application in data scarce scenarios of chemical engineering practice.

physics.chem-ph↗

arXiv2026

Amortized Molecular Optimization via Group Relative Policy Optimization

In structurally constrained molecular optimization, state-of-the-art methods restart an expensive oracle-driven search from scratch for every new input structure, scaling poorly to settings with many starting structures or expensive oracles. While amortized approaches that learn a transferable policy could in principle remove this bottleneck, existing methods struggle to generalize to diverse structural constraints at inference time. We present AMORTIX, an amortized Graph Transformer model that natively supports such constraints, optimizing molecular structures in a single forward pass with zero inference-time oracle calls. A central challenge for amortized training in this domain is that optimization difficulty varies drastically across starting structures. We show that, under this heterogeneity, standard reinforcement learning methods fail to stabilize training, and address this by normalizing rewards within groups of completions sharing the same starting structure. We evaluate on structurally constrained single- and multi-target kinase inhibitor design, and on a few-shot prodrug case study. AMORTIX outperforms both amortized and instance-optimization baselines on goal-directed scaffold decoration and ranks first among amortized methods on the PMO benchmark; the prodrug case study further demonstrates transfer of a learned modification rule to unseen drug structures. Code is available at https://github.com/Hash-hh/AMORTIX/.

cs.LG↗

arXiv2026

Estimating Dense-Packed Zone Height in Liquid-Liquid Separation: A Physics-Informed Neural Network Approach

Separating liquid-liquid dispersions in gravity settlers is critical in chemical, pharmaceutical, and recycling processes. The dense-packed zone height is an important performance and safety indicator but it is often expensive and impractical to measure due to optical limitations. We propose a framework to estimate phase heights by combining a PINN model with readily available volume flow measurements, without requiring phase height measurements during deployment. To this end, a physics-informed neural network (PINN) is first pretrained on synthetic data and physics equations derived from a low-fidelity (approximate) mechanistic model to reduce the need for extensive experimental data. While the mechanistic model is used to generate synthetic training data, only volume balance equations are used in the PINN, as incorporating droplet coalescence and sedimentation submodels would be computationally prohibitive. The pretrained PINN is then fine-tuned with scarce experimental phase height and flow-rate data to capture the actual dynamics of the separator. We then deploy the differentiable PINN as a predictive model in an Extended Kalman Filter inspired state estimation framework, enabling the phase heights to be tracked and updated using flow-rate measurements only. We first test the two-stage trained PINN by forward simulation from a known initial state against the mechanistic model and a non-pretrained PINN. We then evaluate phase height estimation performance with the filter, comparing the two-stage trained PINN with a two-stage trained purely data-driven neural network. All model types are trained and evaluated using ensembles to account for model parameter uncertainty. In all evaluations, the two-stage trained PINN yields the most accurate phase-height estimates.

cs.LG↗

arXiv2026

Data-Driven Conditional Flexibility Index

With the increasing flexibilization of processes, determining robust scheduling decisions has become an important goal. Traditionally, the flexibility index has been used to identify safe operating schedules by approximating the admissible uncertainty region using simple admissible uncertainty sets, such as hypercubes. Presently, available contextual information, such as forecasts, has not been considered to define the admissible uncertainty set when determining the flexibility index. We propose the conditional flexibility index (CFI), which extends the traditional flexibility index in two ways: by learning the parametrized admissible uncertainty set from historical data and by using contextual information to make the admissible uncertainty set conditional. This is achieved using a normalizing flow that learns a bijective mapping from a Gaussian base distribution to the data distribution. The admissible latent uncertainty set is constructed as a hypersphere in the latent space and mapped to the data space. By incorporating contextual information, the CFI provides a more informative estimate of flexibility by defining admissible uncertainty sets in regions that are more likely to be relevant under given conditions. Using an illustrative example, we show that no general statement can be made about data-driven admissible uncertainty sets outperforming simple sets, or conditional sets outperforming unconditional ones. However, both data-driven and conditional admissible uncertainty sets ensure that only regions of the uncertain parameter space containing realizations are considered. We apply the CFI to a security-constrained unit commitment example and demonstrate that the CFI can improve scheduling quality by incorporating temporal information.

cs.LG↗

arXiv2025

End-to-End Reinforcement Learning of Koopman Models for eNMPC of an Air Separation Unit

With our recently proposed method based on reinforcement learning (Mayfrank et al. (2024), Comput. Chem. Eng. 190), Koopman surrogate models can be trained for optimal performance in specific (economic) nonlinear model predictive control ((e)NMPC) applications. So far, our method has exclusively been demonstrated on a small-scale case study. Herein, we show that our method scales well to a more challenging demand response case study built on a large-scale model of a single-product (nitrogen) air separation unit. Across all numerical experiments, we assume observability of only a few realistically measurable plant variables. Compared to a purely system identification-based Koopman eNMPC, which generates small economic savings but frequently violates constraints, our method delivers similar economic performance while avoiding constraint violations.

cs.LG↗

arXiv2025

DeepEOSNet: Capturing the dependency on thermodynamic state in property prediction tasks

We propose a machine learning (ML) architecture to better capture the dependency of thermodynamic properties on the independent states. When predicting state-dependent thermodynamic properties, ML models need to account for both molecular structure and the thermodynamic state, described by independent variables, typically temperature, pressure, and composition. Modern molecular ML models typically include state information by adding it to molecular fingerprint vectors or by embedding explicit (semi-empirical) thermodynamic relations. Here, we propose to rather split the information processing on the molecular structure and the dependency on states into two separate network channels: a graph neural network and a multilayer perceptron, whose output is combined by a dot product. We refer to our approach as DeepEOSNet, as this idea is based on the DeepONet architecture [Lu et al. (2021), Nat. Mach. Intell.]: instead of operators, we learn state dependencies, with the possibility to predict equation of states (EOS). We investigate the predictive performance of DeepEOSNet by means of three case studies, which include the prediction of vapor pressure as a function of temperature, and mixture molar volume as a function of composition, temperature, and pressure. Our results show superior performance of DeepEOSNet for predicting vapor pressure and comparable performance for predicting mixture molar volume compared to state-of-research graph-based thermodynamic prediction models from our earlier works. In fact, we see large potential of DeepEOSNet in cases where data is sparse in the state domain and the output function is structurally similar across different molecules. The concept of DeepEOSNet can easily be transferred to other ML architectures in molecular context, and thus provides a viable option for property prediction.

physics.chem-ph↗

arXiv2025

Molecular Machine Learning in Chemical Process Design

We present a perspective on molecular machine learning (ML) in the field of chemical process engineering. Recently, molecular ML has demonstrated great potential in (i) providing highly accurate predictions for properties of pure components and their mixtures, and (ii) exploring the chemical space for new molecular structures. We review current state-of-the-art molecular ML models and discuss research directions that promise further advancements. This includes ML methods, such as graph neural networks and transformers, which can be further advanced through the incorporation of physicochemical knowledge in a hybrid or physics-informed fashion. Then, we consider leveraging molecular ML at the chemical process scale, which is highly desirable yet rather unexplored. We discuss how molecular ML can be integrated into process design and optimization formulations, promising to accelerate the identification of novel molecules and processes. To this end, it will be essential to create molecule and process design benchmarks and practically validate proposed candidates, possibly in collaboration with the chemical industry.

physics.chem-ph↗

arXiv2025

Accelerating Deterministic Global Optimization via GPU-parallel Interval Arithmetic

Spatial Branch and Bound (B&B) algorithms are widely used for solving nonconvex problems to global optimality, yet they remain computationally expensive. Though some works have been carried out to speed up B&B via CPU parallelization, GPU parallelization is much less explored. In this work, we investigate the design of a spatial B&B algorithm that involves an interval-based GPU-parallel lower bounding solver: The domain of each B&B node is temporarily partitioned into numerous subdomains, then massive GPU parallelism is leveraged to compute interval bounds of the objective function and constraints on each subdomain, using the Mean Value Form. The resulting bounds are tighter than those achieved via regular interval arithmetic without partitioning, but they remain fast to compute. We implement the method into our open-source solver MAiNGO via CUDA in two manners: wrapping all GPU tasks within one kernel function, or distributing the GPU tasks onto a CUDA graph. Numerical experiments show that using more subdomains leads to significantly tighter lower bounds and thus less B&B iterations. Regarding wall clock time, the proposed spatial B&B framework achieves a speedup of three orders of magnitude compared to applying interval arithmetic on the CPU without domain partitioning. Among the two implementations, the one developed with CUDA graph enables higher efficiency. Moreover, in some case studies, the proposed method delivers competitive or better performance compared to MAiNGO's default solver which is based on McCormick relaxations. These results highlight the potential of GPU-accelerated bounding techniques to accelerate B&B algorithms.

math.OC↗

arXiv2025

Simultaneous Design of Microbe and Bioreactor

When developing a biotechnological process, the microorganism is first designed, e.g., using metabolic engineering. Then, the optimum fermentation parameters are determined on a laboratory scale, and lastly, they are transferred to the bioreactor scale. However, this step-by-step approach is costly and time-consuming and may result in suboptimal configurations. Herein, we present the bilevel optimization formulation SimulKnockReactor, which connects bioreactor design with microbial strain design, an extension of our previous formulation, SimulKnock (Ziegler et al., 2024, AIChE J.). At the upper (bioreactor) level, we minimize investment and operation costs for agitation, aeration, and pH control by determining the size and operating conditions of a continuous stirred-tank reactor - without selecting specific devices like the stirrer type. The lower (cellular) level is based on flux balance analysis and implements optimal reaction knockouts predicted by the upper level. Our results with a core and a genome-scale metabolic model of Escherichia coli show that the substrate is the largest cost factor. Our simultaneous approach outperforms a sequential approach using OptKnock. Namely, the knockouts proposed by OptKnock cannot guarantee the required production capacity in all cases considered. In the case that both approaches deliver feasible results, the total annual costs are the same or lower with SimulKnockReactor, highlighting the advantage of combining cellular and bioreactor levels. This work is a further step towards a fully integrated bioprocess design.

math.OC↗

arXiv2025

Nonlinear Model Order Reduction of Dynamical Systems in Process Engineering: Review and Comparison

Computationally cheap yet accurate dynamical models are a key requirement for real-time capable nonlinear optimization and model-based control. When given a computationally expensive high-order prediction model, a reduction to a lower-order simplified model can enable such real-time applications. Herein, we review nonlinear model order reduction methods and provide a comparison of method characteristics. Additionally, we discuss both general-purpose methods and tailored approaches for chemical process systems and we identify similarities and differences between these methods. As machine learning manifold-Galerkin approaches currently do not account for inputs in the construction of the reduced state subspace, we extend these methods to dynamical systems with inputs. In a comparative case study, we apply eight established model order reduction methods to an air separation process model: POD-Galerkin, nonlinear-POD-Galerkin, manifold-Galerkin, dynamic mode decomposition, Koopman theory, manifold learning with latent predictor, compartment modeling, and model aggregation. Herein, we do not investigate hyperreduction, i.e., reduction of floating point operations. Based on our findings, we discuss strengths and weaknesses of the model order reduction methods.

eess.SY↗