ADMOS 2027


Minisymposia

Plenary lectures will be complemented by Minisymposia organised by recognised experts in targeted research areas, relevant for the conference topics.

Participants interested in organising a Minisymposium at ADMOS 2027 are requested to submit an abstract to the Congress Secretariat using the template available here .

Each MS is expected to consist of at least one 2-hour session which may be organised as follows:

- 6 invited talks of 20 minutes;
- 1 keynote lecture of 40 minutes + 4 invited talks of 20 minutes.

The above schedule includes time for questions and discussion. MS failing to fill six presentation slots can be completed by the conference organisers with contributed papers covering topics related to the MS.

Please feel free to contact the Congress Secretariat for any question.

Many tasks in computational science and engineering involve moving beyond forward simulation to quantify uncertainties and solve deterministic or stochastic inverse problems. Such tasks are often characterized as many-query since they typically require multiple evaluations of a forward (defined as mapping parameters to quantities of interest) computational model. When this forward computational model is high-fidelity and expensive to evaluate, it becomes critical to adaptively choose the model evaluations to maximize the information content they are expected to provide. In the context of uncertainty quantification, effective methodologies include both adaptive sampling methods and adaptive construction of surrogate models. In the context of inverse problems, approaches for Bayesian formulations typically seek to concentrate the computational effort in the regions supported by the posterior distribution. In either case, challenges include high-dimensional spaces, noisy or nondeterministic data and/or models, extracting meaningful information from big data, leveraging lower-fidelity information sources, developing goal-oriented approaches, and optimal design of experiments to maximize information content. In this minisymposium, we invite researchers to present recent work on uncertainty quantification and/or inverse problems, with a particular emphasis on robust and efficient adaptive methodologies. Topics include, but are not limited to: adaptive surrogate models, adaptive sampling (including importance sampling) strategies, adaptive sampling for inverse problems (both Bayesian and non-Bayesian), adaptive construction of posterior/updated distributions, adaptive design of experiments, manifold learning, goal-oriented approaches, and active or reinforcement learning.
Organized by: T. Wildey (Sandia National Labs, United States) and A. Belma (Sorbonne University, France)
Keywords: inverse problems, uncertainty quantification
Mathematics of planet earth is a broad term that focuses on the modeling and prediction of a wide range of natural and anthropic-related phenomena, that are often interconnected and whose relevance is hardly underestimated. A non-exhaustive list includes fluid/solid mechanics environmental problems (earthquakes, landslides and rockfalls, erosion, subsidence, groundwater and surface flows, land and sea ice dynamics, sea level rise and saltwater intrusion, tsunamis), weather/climate predictions, and ecological topics such as pollution, vegetation growth (and its counterparts: desertification and wildfires), and population dynamics. All these phenomena are significantly impacted by uncertainties in every step of the mathematical workflow: modeling, prescription of parameters, boundary and initial conditions, geometry and time-scale of the phenomena, measurement errors and data quality in general. Reliably accounting for these uncertainties is key in the decision-making process (planning, hazard control, disaster mitigation). The typical temporal and spatial scales of the phenomena above, their non-linearity and coupled nature, as well as possibly chaotic aspects of the associated dynamics, further exacerbate the theoretical and computational complexities for obtaining meaningful, uncertainty-aware predictions in feasible time. In doing so, adaptivity in its many declinations plays a central role in keeping computational costs affordable. In this MS, we welcome methodological and application-oriented contributions that focus on adaptivity in the treatment of uncertainty: for example, sampling, exploration of the uncertainty domain and active learning, multi-fidelity approaches, construction of surrogate/reduced-order models, data assimilation/filtering and bayesian inversion.
Organized by: C. Piazzola (Universitat Politècnica de Catalunya, Spain), L. Tamellini (CNR - IMATI, Italy) and L. Rinaldi (CNR - IMATI, Italy)
Keywords: adaptivity, mathematics of planet earth, sampling and active learning, surrogate and reduced order modeling, data assimilation and bayesian inversion, multi-fidelity methods, uncertainty quantification
This work investigates a multi-term general fractional advection-diffusion equation, where the time derivative is a general Caputo fractional operator incorporating a weight and a strictly increasing scale function. First, we develop and analyze a numerical scheme. The temporal discretization is implemented on uniform, graded, and adaptive meshes, while a high-order compact difference scheme is employed for the spatial domain. The adaptive mesh is generated through the equidistribution of a positive monitor function. A rigorous stability and convergence analysis is performed for the uniform and graded mesh schemes. We rigorously prove that these schemes achieve an optimal convergence order of $\min\{2-\alpha_1, r\alpha_1\}$ in time and fourth-order accuracy in space, where $\alpha_1$ is the leading fractional order and r is the grading parameter. The performance of the proposed schemes is evaluated through three numerical examples with non-smooth solutions. The adaptive mesh consistently outperforms both graded and uniform meshes in these tests. Furthermore, the results clearly illustrate the significant impact of the weight and scale functions on the solution's behavior.
Organized by: R. Maurya (University of Allahabad, Prayagraj, India)
Keywords: Adaptive mesh, Convergence analysis, Fractional diffusion equation, Numerical analysis, Numerical stability
The global transition toward a low-carbon economy depends critically on two interconnected pillars: modern tools for exploration of critical minerals and clean energy, and green and safe exploitation of the resources. This mini-symposium proposes to address a new generation of tools for the exploration and exploitation in the mining industry. The session is conceived as a collaborative umbrella uniting numerical and technical methodologies in collaboration between academia and industry. Contributions are invited across the full technological spectrum, from the mathematical and computational foundations of deep subsurface exploration — physics-informed artificial intelligence for seismic data reconstruction, adaptive numerical modelling of crustal fluid flows, probabilistic data fusion for geothermal reservoirs, and the use of reduced-order models (ROMs) to accelerate simulations— through to the applied engineering and "Mining 4.0" objectives of modern exploitation, including the deployment of real-time monitoring tools and computational control systems in active operations that enhance efficiency and provide early warnings. By treating exploration and exploitation as a continuum rather than as separate concerns, the mini-symposium aims to make explicit the shared mathematical machinery that underpins both, and to accelerate the transfer of advanced computational methods from scientific research to industrial practice. The proposed format is intended to bring together international experts from academia and industry, encouraging the emergence of joint research directions and collaborative initiatives.
Organized by: A. García-González (UPC/CIMNE, Spain), A. Muixí (UPC/CIMNE, Spain), I. Markovsky (ICREA/CIMNE, Spain) and S. Zlotnik (UPC/CIMNE, Spain)
Keywords: adaptive numerical modelling, computational control systems, IoT, mathematics of planet earth, Mining 4.0, physics-informed artificial intelligence , real-time monitoring, surrogate and reduced order modeling
The use of experimental data in combination with simulation models has become a very active research topic. Indeed, new sensing facilities (such as those related to image-based or optic fiber-based technologies) now enable to collect a large and diversified amount of data from a physical system. Such data may be used to identify and validate complex models, or to build hybrid models in which machine learning components enhance the predictions of a physics-based simulator. This represents the “physical-to-virtual” leg of the digital twin framework. However, many challenges dealing with data selection/filtering, uncertainty quantification, management of computational cost, or numerical robustness, need to be addressed in order to incorporate data efficiently in practical applications. The goal of this mini-symposium is to bring together recent fundamental advances in data assimilation, inverse methods, and hybrid modeling, in both deterministic and stochastic (Bayesian) settings. With regards to innovative and powerful numerical approaches which emerged recently, we anticipate contributions on the following topics: - use of model reduction or adaptive/multi-fidelity strategies for high dimensional parameter/state spaces; - offline operator/surrogate learning, subsequently updated online, versus all-at-once formulations in which state and parameters are inferred simultaneously; - real-time sequential model updating for DDDAS (Dynamic Data-Driven Application Systems), requiring fast inference algorithms; - physics-informed and physics-augmented algorithms for deep learning from data; - analysis of full-field measurements and large data sets, together with the representation and propagation of modeling and measurement errors from the raw signal up to the identified quantities; - goal-oriented (“fit-for-purpose”) model updating; - optimal experimental design for informative data acquisition; - applications in multiscale or multi-physics contexts.
Organized by: M. Beer (Leibniz University Hannover, Germany), R. Bouclier (INSA Toulouse, University of Toulouse, France), L. Chamoin (ENS Paris-Saclay, France), E. Chatzi (ETH Zurich, Switzerland) and E. Cueto (University of Zaragoza, Spain)
Keywords: adaptivity, data assimilation and bayesian inversion, inverse problems, physics-informed artificial intelligence, surrogate and reduced order modeling, uncertainty quantification
MESH ADAPTATION FOR FLOW SIMULATION A. GARGALLO-PEIRÓ, J. DOCAMPO-SÁNCHEZ AND O. LEHMKUHL Barcelona Supercomputing Center Plaça Eusebi Güell, 1-3 08034 Barcelona (Spain) ABSTRACT The demand for advanced Computational Fluid Dynamics (CFD) tools to meet modern engineering design challenges continues to grow rapidly. In industries such as aerospace, automotive, and energy, there is a pressing need to re-engineer CFD solvers to fully exploit the potential of extreme-scale parallel computing platforms. This mini-symposium on Mesh Adaptation for Flow Simulation will gather experts to explore recent breakthroughs in mesh refinement, metric-based adaptation, and remeshing, with a special emphasis on industrial applications like wind resource assessment or high-lift aerodynamics for aircraft design. Key topics include: 1. Adaptive Mesh Refinement (AMR) and Metric-Based Adaptation: Addressing challenges in dynamically refining, coarsening, and aligning anisotropic meshes (e.g., Hessian/error-metric driven, octree-based AMR) to improve CFD accuracy and efficiency. Focus includes applications across RANS, Scale-Resolving Simulations (LES/DES), and complex geometries. 2. Mesh Adaptation for High-Order Schemes: Advancements tailored for low-dissipation, high-order discretization methods (p-refinement, Discontinuous Galerkin, high-order FV/FE). These methods enhance accuracy in both incompressible and compressible flows, particularly in turbulent boundary layers and aerodynamics. 3. Data-Driven & AI-Guided Adaptation: Machine learning and AI frameworks used to steer mesh adaptation in CFD. Topics include feature detection via statistical methods, reinforcement learning for optimal mesh distribution, and data-driven turbulence modeling. 4. Industrial Case Studies & HPC Scalability: Real-world applications of large-scale adaptive simulations across aeronautical, automotive, and wind energy sectors, highlighting execution on extreme-scale computing architectures. The symposium aims to advance the conversation on mesh adaptation by bringing together innovative numerical techniques, HPC applications, and industrial case studies.
Organized by: A. Gargallo-Peiró (BSC-CNS, Spain), J. Docampo-Sánchez (BSC-CNS, Spain) and O. Lehmkuhl (BSC-CNS, Spain)
Keywords: Adaptive mesh, adaptivity, multi-fidelity methods
Error estimation and adaptation are essentiel for efficient and reliable computer simulations of complex phenomena. Indeed, the numerical solution of many practical problems in computational sciences and engineering is challenging, and designing adaptive methods is of paramount importance for reducing computational costs while retaining a suitable level of accuracy. The topic now extends beyond classical discretization error assessment and mesh refinement as in standard hp adaptive finite element methods as it also encompasses adaptive reduced-order modeling for building accurate surrogate models of random/parametric partial differential equations, models which are particularly useful in the context of multi-query simulations for uncertainty quantification, optimization, or inverse design. It further includes novel topics relevant to engineering applications, including goal-oriented methods for the control of errors in specific quantities of interest. The objective of the mini-symposium is to present the latest developments on goal- oriented error estimation and adaptive methods. We anticipate talks on topics including: • Discretization and modeling error analysis for nonlinear or time-dependent problems; • Control of errors in multi-physics or multi-scale problems; • r-adaptive methods based on moving mesh or optimal transport techniques; • Adaptive strategies for reduced-order models; • Error estimation and adaptive schemes for uncertainty quantification; • Stability, convergence, and optimality analysis of adaptive methods; • Error estimation for scientific machine learning; • Use of machine learning techniques for error estimation and adaptivity.
Organized by: D. Guignard (University of Ottawa, Canada), K. Kergrene (Université de Technologie de Compiègne, France) and S. Prudhomme (Polytechnique Montréal, Canada)
Keywords: Adaptive Modeling, Adjoint Problem, Convergence Analysis , Goal-Oriented Error Estimation, Quantities of Interest, Reduced Order Modeling, Scientific Machine Learning
All numerical simulations are based on the combination of a physical model for the underlying system behaviour, and a numerical discretisation of this model. But while adaptive methodologies for the numerical part of the simulation, based on sophisticated error estimations, have reached a high degree of maturity, adaptive methods for physical models are not yet encountered frequently. Instead, the same physical model tends to be used throughout a simulation, the choice of model depends on the experience of the user, and uncertainties are approximated through comparisons with similar simulations that have been validated with experiments. Potential difficulties in creating locally adaptive physical models are, first, that no suitable hierarchies of models, from cheap to accurate, may be available. A second challenge is to define consistent couplings between these models, to allow each one to be applied locally. And finally, error models for physical models (equivalent to the well-known power-law behaviour of many numerical discretisations) are not available. Nevertheless, for example the recent advances in machine learning could bring such models within reach. The goal of this exploratory minisymposium is to gather the participants’ thoughts on the feasibility of solution-adaptive physical models. Topics include, but are not limited to, model coupling and hybridisation, adaptive microscale models, data-driven models, combined numerical and physical adaptation, model selection criteria, and error indicators. Applications could include turbulence modelling, boundary and shear layers, nonhomogeneous materials, damage growth, etc. Early-stage research, incomplete work, or even conceptual discussions are very much encouraged. It is hoped that these discussions will contribute to the evolution of reliable fully automated simulation of physical phenomena.
Organized by: J. Wackers (Centrale Nantes / CNRS, France)
Keywords: adaptivity, error estimation , physical models, uncertainty quantification
The increasing complexity of computational models in engineering and science demands numerical methods that can efficiently resolve multiscale and evolving physical phenomena while controlling computational cost. Mesh adaptation provides a fundamental mechanism for achieving this by dynamically allocating computational resources where they are most needed. This mini-symposium will bring together researchers working on h-, p-, and r-adaptation, covering both low- and high-order discretisations, mesh motion, and emerging data-driven approaches for adaptive simulation.
The symposium will address recent advances in adaptive mesh generation and simulation, including anisotropic and goal-oriented adaptation, high-order and curved meshes, dynamic mesh motion, and combined h/p/r strategies. Particular attention will be given to challenges arising in large-scale and high-performance computing, including parallel mesh adaptation, load balancing, scalability to very large meshes, and adaptation for complex geometries and multiphysics applications.
A further focus will be the emerging role of artificial intelligence and machine learning in mesh adaptation. Topics of interest include learning-based error and feature indicators, data-driven prediction of optimal mesh resolution and polynomial degree, reinforcement learning for adaptive strategies, transfer learning between simulations, and physics-informed or physics-aware approaches. The symposium will explore how AI can complement established adaptation techniques while addressing challenges of robustness, interpretability, computational cost, and generalisation.
Contributions are invited that advance the state of the art in theory, algorithms, software, and applications of mesh adaptation. The symposium aims to provide a forum for interaction between researchers developing numerical methods, mesh-generation technologies, high-performance computing approaches, and AI-based techniques, with the goal of identifying new directions towards automated, efficient, scalable, and intelligent adaptive computational simulations.
Organized by: R. Sevilla (Swansea University, United Kingdom), D. Moxey (King’s College London, United Kingdom), J. Peiró (Imperial College London, United Kingdom) and O. Hassan (Swansea University, United Kingdom)
Keywords: Adaptive Modeling, adaptivity
Engineering systems must increasingly combine multiple functions, operate across changing conditions, and maintain or adapt their performance in the presence of uncertainty. Topology optimisation is thus extending beyond conventional performance-driven design towards engineering systems in which geometry and material distribution are tailored to achieve specific functions and tasks. Topology therefore becomes a key determinant of how function is achieved: how energy is harvested, heat is managed, waves are guided, motion or shape change is generated, or interactions with physical and biological environments take place. These functions often emerge from the interplay of several physical phenomena and may require different, even competing, topological features across operating states. This poses new challenges for topology optimisation, requiring methods that can efficiently address parameter-dependent design problems, coupled multi-physics phenomena, evolving geometries, uncertain operating conditions, and high-dimensional design spaces. Advances in numerical algorithms, adaptivity, error estimation, high-performance computing and data-driven modelling offer new opportunities to improve the accuracy, reliability and computational efficiency of the optimisation process. This mini-symposium aims to bring together researchers advancing topology optimisation methods and applications, emphasising the interplay between methodological innovation and emerging functional requirements. Methodological topics of interest include: adaptivity and error estimation; multi-resolution and multi-scale optimisation; parallel and GPU computing; multigrid and preconditioning; surrogate and multi-fidelity modelling; uncertainty quantification and robust optimisation; AI-assisted optimisation and generative models; quantum and hybrid quantum-classical approaches; CAD-compatible and manufacturing-aware formulations. Applications of interest include: multi-scale and multi-functional metamaterials; energy harvesting; thermal-fluid systems and battery thermal management; programmable and adaptive materials; active, shape-morphing, and reconfigurable devices; soft robotics; coupled actuation-control design; acoustic, photonic, and electromagnetic devices; vibration and noise control; bio-integrated, biohybrid, and tissue-interacting systems.
Organized by: M. Giacomini (Universitat Politècnica de Catalunya - CIMNE, Spain), J. Principe (Universitat Politècnica de Catalunya - CIMNE, Spain) and A. Ferrer (Universitat Politècnica de Catalunya - CIMNE, Spain)
Keywords: adaptive systems, generative models, parametric problems, quantum computing, robust optimization, topology optimization
Many problems in engineering and science involve solving problems on multiple time- and length-scales. Numerous techniques have been developed with the aim to concurrently solve for the relevant physics on two or more length-scales. Examples of such techniques are the Homogenization techniques (Computational homogenization, “finite element square”, etc.), subgrid techniques (the Variational MultiScale method, Multi-Scale FEM, Localized Orthogonal Decomposition, etc,) and other coarse graining techniques (e.g. the Quasi-Continuum Method) to name a few examples. In common for most of these procedures is the solution of a local problem. For data stemming from the higher scale, a fine-scale problem is solved for, and effective outputs are returned to the higher scale. Such implicit procedures, with a high number of evaluations of the fine-scale models, results in a high computational cost.
To make on-the-fly multi-scale analysis feasible for practical problems, reduced order modeling has become a popular means to reduce the cost of the fine-scale problems. Reduced order models (or surrogates) can either be constructed based purely on input-output data, or in terms of reducing the problem size. It is imperative to control the accuracy of such procedures, and to balance the accuracy of the overall computation against the computational cost.
The purpose of this mini symposium is to disseminate and discuss recent advances in reduced order modeling approaches (primarily) within the scope of multiscale modeling. More specifically, such procedures will be discussed in view of balancing the accuracy vs. computational effort. Topics include, but are not limited to
  • Adaptive multiscale methods
  • Error estimation for multiscale-problems and/or reduced order models
  • Novel ROM techniques
  • Application of established ROM techniques to novel applications (e.g., coupled problems)
  • Adaptive basis enrichment
  • Adaptive quadrature (Hyper-reduction)
  • Training of reduced models (for, e.g., Neural Networks or POD-based models)
Organized by: F. Larsson (Chalmers, Sweden) and R. Jänicke (TU Braunschweig, Germany)