Numerical Methods Scientific Computing

Physics-Informed Neural Networks

Development of neural network architectures that incorporate physical conservation laws and differential equations as constraints for solving complex scientific computing problems.

RESEARCH GAP FRONTIERS

Uncertainty Quantification in Computational Models

Methods for characterizing and propagating uncertainties through numerical simulations to provide reliable confidence bounds on scientific predictions.

RESEARCH GAP FRONTIERS

Machine Learning-Enhanced Surrogate Modeling

Construction of reduced-order models using machine learning to approximate expensive computational simulations with minimal loss of accuracy.

RESEARCH GAP FRONTIERS

Adaptive Mesh Refinement Algorithms

Computational techniques that dynamically adjust spatial discretization based on solution features to optimize accuracy and computational efficiency.

RESEARCH GAP FRONTIERS

High-Order Discontinuous Galerkin Methods

Advanced finite element techniques allowing solutions with jumps across element boundaries to achieve superior accuracy for hyperbolic systems.

RESEARCH GAP FRONTIERS

Spectral Element Methods for Complex Geometries

High-accuracy numerical schemes combining spectral methods with element-based decomposition for efficient solution of PDEs on intricate domains.

RESEARCH GAP FRONTIERS

Fast Multipole Method Development

Algorithmic innovations to accelerate computation of long-range interactions in N-body problems reducing complexity from O(N²) to O(N).

RESEARCH GAP FRONTIERS

Implicit-Explicit Time Integration Schemes

Hybrid temporal discretization methods that treat stiff and non-stiff components separately to achieve stability and computational efficiency.

RESEARCH GAP FRONTIERS

Multiscale Numerical Homogenization

Techniques for bridging disparate spatial and temporal scales in heterogeneous materials to obtain effective macroscopic properties.

Isogeometric Analysis Framework

Integration of computer-aided design representations with numerical analysis to maintain geometric accuracy throughout the solution process.

Stabilized Finite Element Methods

Variational formulations with additional stabilization terms to suppress spurious numerical oscillations in convection-dominated problems.

Preconditioned Krylov Subspace Solvers

Development of advanced conditioning techniques and iterative methods to accelerate convergence of large sparse linear systems.

Multigrid Methods for Heterogeneous Media

Hierarchical computational approaches designed to efficiently solve equations in materials with rapidly varying coefficients.

Boundary Element Method Advances

Improvements in integral equation formulations and fast algorithms for problems where boundary-only discretization is advantageous.

Finite Volume Methods for Nonlinear Conservation Laws

Numerical schemes guaranteeing conservation properties while handling shocks and discontinuities in hyperbolic systems.

Domain Decomposition Methods Optimization

Algorithms partitioning computational domains into subproblems for parallel processing while maintaining solution continuity.

Variational Data Assimilation Techniques

Methods for optimally combining observational data with computational models to improve predictions in dynamical systems.

Gradient-Enhanced Kriging for Design Optimization

Surrogate-based optimization using derivative information to improve metamodel accuracy for expensive objective functions.

Oscillatory Integral Computation Methods

Specialized quadrature and asymptotic techniques for accurately integrating rapidly oscillating functions in wave-based applications.

Exponential Integrator Development

Time stepping schemes based on exponential matrix functions for accurate simulation of stiff differential equations.

Meshfree Methods with Radial Basis Functions

Computational approaches avoiding traditional meshes by using radial basis function interpolation for flexible spatial discretization.

Discontinuous Petrov-Galerkin Methods

Advanced weak formulations providing automatic stability without explicit stabilization for diverse PDE systems.

Flux Reconstruction Methods Development

High-order numerical schemes unifying multiple formulations through common flux reconstruction framework for efficient computation.

Iterative Refinement for Ill-Posed Problems

Regularization techniques using iterative procedures with early stopping to solve ill-conditioned inverse problems stably.

Quantum Computing Algorithms for Linear Systems

Development of quantum algorithms providing potential speedups for solving sparse linear systems in scientific computing.

Deep Learning for PDE Solution Approximation

Neural network-based approaches for directly approximating PDE solutions with ability to generalize across parameter spaces.

Tensor Decomposition for High-Dimensional Problems

Efficient representations of multivariate data using low-rank tensor formats to combat curse of dimensionality.

GPU-Accelerated Numerical Algorithms

Algorithmic adaptations and implementations optimized for massive parallelism on graphics processing units.

Collocation Methods for Boundary Value Problems

Point-based discretization schemes enforcing solution accuracy at selected nodes for efficient solving of multi-point problems.

Moving Mesh Methods with Monitor Functions

Techniques where mesh nodes dynamically reposition based on solution gradients to concentrate resolution where needed.

Rigorous Error Analysis for Finite Elements

Mathematical frameworks providing guaranteed error bounds and convergence rates for finite element discretizations.

Polynomial Chaos Expansion Uncertainty Analysis

Orthogonal polynomial-based representations of random solutions enabling efficient uncertainty propagation in stochastic PDEs.

Artificial Viscosity Methods for Shocks

Numerical dissipation techniques preventing oscillations near discontinuities while maintaining solution smoothness in regular regions.

Reduced Basis Method Certification

A posteriori error estimators for reduced-order models guaranteeing solution quality for parametric PDE families.

Coupled Multiphysics Simulation Methods

Algorithms for solving interconnected systems of PDEs from different physics domains with appropriate coupling strategies.

Hermite-Birkhoff Interpolation Schemes

Advanced interpolation using derivative information at multiple points to improve accuracy of function approximations.

Stabilized Mixed Finite Element Formulations

Variational methods treating multiple unknowns simultaneously with stabilization to satisfy discrete inf-sup conditions.

Automatic Differentiation for Scientific Computing

Computational techniques computing exact derivatives of complex algorithms for optimization and sensitivity analysis.

Graph Neural Networks for PDE Solving

Machine learning architectures leveraging graph structures to learn solution operators for PDEs on arbitrary geometries.

Conservative Numerical Schemes for Hamiltonian Systems

Time integrators preserving energy and symplectic structure for long-term accurate simulation of conservative dynamics.

Parametric Sensitivity Analysis Methods

Efficient techniques for computing derivatives of model outputs with respect to input parameters for design optimization.

Weighted Essentially Non-Oscillatory Schemes

High-order methods achieving non-oscillatory solutions near discontinuities through adaptive weighting of stencil contributions.

Quasi-Monte Carlo Methods for Integration

Low-discrepancy sequence-based quadrature achieving faster convergence than traditional Monte Carlo for high-dimensional integrals.

Finite Difference Schemes for Fractional Derivatives

Discrete approximations for non-integer order derivatives enabling accurate numerical solution of fractional differential equations.

Operator Splitting Methods for Complex PDEs

Techniques decomposing complex PDE systems into simpler subproblems solved sequentially or in parallel.

Structure-Preserving Schemes for Optimization

Numerical algorithms maintaining essential mathematical properties like convexity during iterative solution of optimization problems.

Neural Operator Learning Theory

Theoretical foundations for training neural networks to learn infinite-dimensional operators mapping between function spaces.

Continuous Remap for Multi-Material Flows

Algorithms for accurately tracking material interfaces and remapping quantities between Eulerian and Lagrangian representations.

Localized Orthogonal Decomposition Method

Multiscale finite element technique using local problems to construct basis functions capturing fine-scale heterogeneity.

Ensemble Kalman Filter for Data Assimilation

Particle-based sequential estimation method combining measurements with model predictions in nonlinear filtering problems.

Virtual Element Methods for Polygonal Meshes

Development and analysis of VEM schemes for arbitrary polygonal and polyhedral discretizations with applications to complex domain geometries.

Structure-Preserving Symplectic Integrators

Design of time integration methods that preserve Hamiltonian structure and symplectic geometry in molecular dynamics and celestial mechanics.

Hybridizable Discontinuous Galerkin Methods

Development of HDG formulations enabling efficient static condensation and reduced computational complexity for multiphysics problems.

Localized Radial Basis Function Collocation

Local RBF approximation techniques with compactly supported kernels for scattered data interpolation and PDE solution.

Entropy-Stable Finite Volume Schemes

Design of numerical methods satisfying discrete entropy inequalities for hyperbolic conservation laws and compressible flow.

Implicit-Explicit Runge-Kutta Methods

Analysis and optimization of IMEX-RK schemes for stiff-nonstiff partitioned systems in multiscale dynamics.

Least Squares Finite Element Methods

First-order system least-squares formulations providing symmetric positive-definite systems for coupled PDE problems.

Compatible Finite Element Discretizations

Mimetic and compatible FEM pairs preserving commuting diagram properties for electromagnetics and fluid dynamics.

Fractional-Step Methods for Incompressible Flow

Projection and splitting methods decoupling pressure from velocity in Navier-Stokes equations with temporal accuracy analysis.

Sparse Grid Stochastic Collocation

High-dimensional uncertainty quantification using Smolyak sparse grids and tensor product collocation for parametric PDEs.

High-Order Accurate Shock Capturing

Development of WENO-type and flux limiting techniques maintaining high accuracy away from discontinuities in hyperbolic systems.

Oversampling Domain Decomposition Methods

Two-level FETI and BDDC preconditioners with oversampling extensions for scalable parallel linear system solvers.

Perfectly Matched Layer Absorbing Boundaries

Design and analysis of PML formulations for wave equation truncation on unbounded domains in acoustics and electromagnetics.

Conservation Laws with Source Terms

Well-balanced numerical schemes maintaining equilibrium solutions for balance laws in shallow water and gravitation.

Nonoverlapping Schwarz Methods with Transmission Conditions

Optimized boundary conditions for non-overlapping domain decomposition accelerating convergence in parallel computing.

Integral Equation Methods for Helmholtz

Boundary integral and coupled FEM-BEM formulations for exterior scattering problems with wavenumber dependence.

Adaptive Time Stepping Control Strategies

Error-based and embedded Runge-Kutta pair methods for automatic time step selection in transient simulations.

Generalized Minimal Residual GMRES Variants

Restarted, deflated, and recycled GMRES methods with flexible preconditioning for large-scale nonsymmetric systems.

Convolution Quadrature for Time Domain

Discretization of time-domain convolution integrals for time-dependent integral equations in transient wave propagation.

Reconstruction-Based Discontinuous Galerkin

Flux reconstruction and hybrid high-order methods combining DG flexibility with efficient matrix assembly.

Certified Reduced Order Model Error Bounds

Rigorous a posteriori error estimation and certification for parametric reduced basis approximations of PDEs.

Multigrid Methods for Nonlinear Problems

Full approximation storage and nonlinear multigrid strategies for efficient solution of nonlinear systems.

Local Discontinuous Galerkin Methods

LDG formulations for diffusive PDEs enabling high-order accuracy with compactly supported numerical fluxes.

Block Kaczmarz Iterative Methods

Randomized and deterministic block iterative methods for large-scale overdetermined and underdetermined systems.

Finite Element Methods for Maxwell Equations

Nedelec edge elements and mixed formulations preserving divergence constraints in electromagnetic wave propagation.

Continuous and Discontinuous Petrov-Galerkin

Optimal test function spaces achieving minimum energy extensions for robust and pressure-robust flow problems.

Time Integration with Variable Step Sizes

Analysis and implementation of Adams and backward differentiation formulas with adaptive step size control.

Spectral Vanishing Viscosity Methods

Mollification techniques for spectral methods suppressing Gibbs oscillations near shocks in hyperbolic systems.

Prolongation Restriction Operators Optimization

Design of efficient transfer operators for multigrid methods on unstructured and adaptive meshes.

Enriched Galerkin Methods for Multiscale

Multiscale basis enrichment combining global and local problems for heterogeneous coefficient PDEs.

Finite Element Exterior Calculus

Mimetic discretization of differential forms preserving Hodge-de Rham complex sequences in computational physics.

Block Preconditioning for Saddle Points

Two-by-two block preconditioners for mixed finite element systems arising in fluid and elasticity problems.

Parareal Algorithm for Time Parallelization

Parallel-in-time integration combining coarse and fine solvers for breaking sequential time stepping bottlenecks.

Variational Physics-Informed Neural Operators

Combining variational formulations with neural operator learning for parameterized PDE solution mapping.

Duality-Based Error Estimation Methods

Adjoint-based a posteriori error bounds for goal-oriented adaptivity in finite element approximation.

Mortar Finite Element Coupling

Non-conforming interface conditions for coupling different discretizations and material domains with flexible mesh alignment.

Discontinuous Galerkin Schemes for Kinetic Equations

DG discretization of Boltzmann and Vlasov equations with entropy stability and asymptotic preserving properties.

Fast Transform Methods for Structured Grids

FFT-based and hierarchical matrix techniques for rapid evaluation on regular and tensor-product domains.

Localization in Domain Decomposition

Localized orthogonal decomposition and multiscale domain decomposition for problems with multiscale coefficients.

Positivity Preserving Numerical Schemes

Design of methods maintaining physical constraints like positivity in density and energy variables.

Legendre-Tau Spectral Methods

Spectral collocation using Legendre polynomials with tau correction for high-accuracy differential equation solution.

Schur Complement Reduction Techniques

Static condensation and substructuring methods for efficient solution of large-scale finite element systems.

Numerical Methods for Nonlocal Equations

Finite element and collocation schemes for fractional and peridynamic PDEs with long-range interactions.

Stabilization Techniques for Convection-Dominated Flow

Streamline upwind Petrov-Galerkin and local projection stabilization preventing spurious oscillations.

Monte Carlo Methods for Integration

Quasi-Monte Carlo and importance sampling techniques for high-dimensional integration in uncertainty quantification.

Compact Finite Difference Approximations

Schemes providing higher accuracy with smaller stencils through implicit finite difference formulations.

Primal Dual Active Set Algorithms

Interior point and semi-smooth Newton methods for constrained optimization in PDE-constrained problems.

Mixed Virtual Element Methods

VEM formulations for mixed systems maintaining discrete stability on arbitrary polygonal partitions.

Backward Error Analysis for ODEs

Modified equations and backward error investigation characterizing accuracy and long-term behavior of numerical integrators.

Compressible Flow Limiting Strategies

Slope limiters and discontinuity detection for high-order schemes in compressible Euler and Navier-Stokes equations.

Virtual Element Method Extensions

Research on polygonal and polyhedral finite elements with applications to complex geometries and non-conforming mesh structures.

Exponential Rosenbrock Methods

Development of exponential integrators combining Rosenbrock stages for stiff nonlinear PDEs with improved stability properties.

Fractional Differential Equations Numerics

Numerical schemes for anomalous diffusion and non-integer order derivative operators in scientific computing applications.

Discontinuous Galerkin-Collocation Hybrids

Hybrid formulations combining discontinuous Galerkin and collocation point techniques for improved computational efficiency.

Machine Learning Accelerated PDE Solvers

Integration of neural networks with classical numerical methods to accelerate convergence and reduce computational cost.

Mimetic Finite Difference Methods

Development of finite difference schemes preserving mathematical structure and conservation laws of continuous PDEs.

High-Order Time Integration for Wave Equations

Construction and analysis of symplectic and energy-stable time stepping schemes for hyperbolic wave propagation problems.

Surrogate-Assisted Gradient Estimation

Methods for accurate gradient computation using reduced-order models in inverse problems and optimization contexts.

Structure-Preserving Schemes for Shallow Water

Well-balanced numerical methods maintaining lake-at-rest equilibrium and gravitational equilibrium for shallow water equations.

Partitioned-Monolithic Coupling Strategies

Advanced time integration schemes for coupled multi-physics problems balancing stability and computational efficiency.

Fourier Neural Operators Framework

Research on spectral neural operators for learning solution operators of parametric PDEs with spectral efficiency.

Volumetric Model Order Reduction

Techniques for constructing reduced-order models of three-dimensional problems with guaranteed accuracy certificates.

Positivity-Preserving Numerical Methods

Schemes maintaining physically meaningful non-negativity constraints in density, concentration, and population models.

Conformal Mapping in Numerical Analysis

Application of conformal transformations for mesh generation and solution of elliptic problems in complicated domains.

Transient Stability Analysis Methods

Numerical techniques for analyzing time-dependent stability in power systems and electrical network simulations.

Localized Orthogonal Decomposition Extensions

Advanced localization techniques for model reduction in heterogeneous media with multi-scale features.

Thermodynamically Consistent Discretizations

Numerical schemes satisfying second law of thermodynamics and entropy stability for complex fluid flows.

Adjoint-Based Sensitivity via Automatic Differentiation

Systematic computation of sensitivities using reverse-mode automatic differentiation for inverse problems.

Finite Element Homogenization Theory

Rigorous mathematical frameworks for upscaling heterogeneous materials to effective macroscopic properties.

Conservative Semi-Lagrangian Schemes

Mass-conserving semi-Lagrangian methods for advection-dominated problems with large time steps.

Optimization via Learned Surrogates

Derivative-free optimization frameworks leveraging neural network surrogates for expensive function evaluations.

Rational Krylov Subspace Methods

Development of rational Krylov techniques for solving nonlinear matrix equations and frequency-dependent problems.

Wavelet-Based Numerical Schemes

Construction of adaptive wavelet methods for multi-resolution representation of solutions with compression properties.

Entropy-Stable Formulations for Systems

Provably entropy-dissipative numerical schemes for hyperbolic conservation laws satisfying thermodynamic constraints.

Cartesian Grid Immersed Boundary Methods

Techniques for solving PDEs on fixed Cartesian grids with moving or complex boundaries via penalty methods.

Sparse Grid Approximation Theory

Analysis and development of sparse grid collocation methods for high-dimensional parametric uncertainty problems.

Neural Network Finite Elements

Integration of neural network basis functions within finite element frameworks for improved approximation.

Stochastic Gradient Methods for PDE Solving

First-order optimization algorithms for solving large-scale systems arising from PDE discretizations.

Hybrid Discontinuous Galerkin-FEM

Coupled formulations using discontinuous and continuous elements in different domain regions for efficiency.

Conservation Law Positivity Verification

Formal verification methods ensuring numerical schemes maintain physical constraints throughout simulations.

Graph-Based Domain Decomposition

Application of graph partitioning and spectral methods for optimal subdomain division in parallel computing.

Implicit-Explicit Runge-Kutta Schemes

Development of additive Runge-Kutta methods separating stiff and non-stiff components for efficiency.

Magnetic Field Constrained Solving

Specialized numerical methods maintaining divergence-free constraints in electromagnetic simulations.

Learned Optimal Coefficients Discovery

Data-driven approaches for discovering optimal parameters in numerical schemes using machine learning.

Coupled Flow-Transport Numerical Analysis

Numerical techniques for strongly coupled groundwater flow and contaminant transport simulations.

Metamodel-Based Inverse Problem Solving

Use of calibrated surrogate models as forward operators in Bayesian inference and inverse problems.

Heterogeneous Preconditioner Development

Preconditioners exploiting heterogeneous computing resources including CPUs and GPUs for efficient solves.

Multirate Time Integration Methods

Schemes using different time step sizes for fast and slow physics to improve overall computational efficiency.

Legendre Polynomial Spectral Methods

High-order accurate spectral methods using Legendre polynomial bases for smooth solution approximation.

Coupled Geomechanics-Fluid Flow

Numerical techniques for poroelasticity problems combining deformation and fluid migration in porous media.

Adaptive Finite Volume Schemes

Finite volume methods with automatic mesh adaptation based on solution error indicators and gradients.

Lattice Boltzmann Method Applications

Development and analysis of lattice Boltzmann schemes for fluid dynamics and kinetic transport phenomena.

Constraint-Preserving Evolution Equations

Methods maintaining constraint manifolds in systems with algebraic-differential equation coupling.

High-Dimensional Function Approximation

Techniques for approximating high-dimensional functions using manifold learning and tensor networks.

Staggered Grid Finite Difference Methods

Offset grid schemes for coupled systems improving discrete divergence stability and wave dispersion.

Regression-Based Model Order Reduction

Non-intrusive reduced-order modeling using supervised learning from high-fidelity simulation snapshots.

Isogeometric Collocation Methods

Isogeometric analysis combined with collocation points for efficient high-order PDE discretizations.

Multisymplectic Spatial Discretizations

Discretizations preserving multisymplectic structure for Hamiltonian PDEs ensuring long-time stability.

Convolution Neural Networks for Image Reconstruction

Deep learning approaches for inverse problems and tomographic image reconstruction using convolutional architectures.

Virtual Element Methods

Novel discretization techniques allowing arbitrary polygonal/polyhedral elements with applications to general purpose PDEs.

Stochastic Collocation Methods

Non-intrusive uncertainty propagation via sparse grid collocation points for parametric PDE problems.

Localized Model Order Reduction

Construction of problem-dependent basis functions for efficient reduced order modeling of transport-dominated systems.

Spectral Methods for Unbounded Domains

Development of mapped Chebyshev and Laguerre spectral techniques for computations on infinite intervals and half-lines.

Phase-Field Numerical Simulations

Computational methods for diffuse interface models capturing interfacial phenomena in multiphase flows and materials.

Asymptotic-Preserving Schemes Design

Numerical schemes maintaining correct behavior across different parameter regimes without resolving thin layers.

Higher-Order Time Integration

Development of high-stage Runge-Kutta and multi-step methods with improved stability regions for stiff systems.

Lattice Boltzmann Method Advances

Enhanced kinetic schemes for fluid dynamics with non-equilibrium distribution functions and collision operators.

Parametric Reduced Basis Methods

Greedy construction of parameter-dependent reduced order models with rigorous a posteriori error bounds.

Hybrid High-Order Methods

Bridging cell-based and face-based unknowns for improved flexibility in handling geometric complexities.

Nonlocal Differential Equation Modeling

Numerical treatment of integro-differential equations arising in peridynamics and anomalous transport phenomena.

Inverse Problem Regularization Theory

Tikhonov and Bayesian regularization frameworks for reconstructing unknown parameters from noisy measurements.

Optimized Schwarz Methods

Development of overlapping and non-overlapping domain decomposition schemes with optimized transmission conditions.

Structure-Preserving Integration

Symplectic and energy-stable time integrators for Hamiltonian and dissipative systems with long-time accuracy.

Radial Basis Function Collocation

Meshfree approximation techniques using RBF interpolation for irregular domain geometries and scattered data.

Coupled Fluid-Structure Interaction

Monolithic and partitioned schemes for coupled multiphysics simulations of deformable structures in flows.

Heterogeneous Multiscale Methods

Framework for coupling macroscopic solvers with microscale simulations to capture effective behavior.

Enriched Finite Element Methods

Local basis function enrichment for capturing singularities and localized phenomena without mesh refinement.

Implicit-Explicit Operator Splitting

Decomposition strategies treating stiff and nonstiff terms separately for improved computational efficiency.

Compatible Discretization Schemes

Mixed finite elements satisfying discrete versions of Hilbert complex properties for vector PDEs.

Nonlinear Eigenvalue Problems

Iterative solution methods for parameter-dependent eigenvalue equations arising in photonics and acoustics.

Optimization-Based Model Reduction

Least-squares and optimization frameworks for constructing interpretable surrogate models from high-fidelity data.

Shock-Capturing Finite Element Methods

Entropy-stable and residual-based stabilization for resolving sharp discontinuities in conservation laws.

Generalized Finite Element Methods

Partition of unity enrichment strategies for handling cracks, inclusions, and material interfaces.

Symplectic Runge-Kutta Methods

Construction and analysis of Runge-Kutta schemes preserving symplectic structure for canonical Hamiltonian systems.

Finite Element Method for Eigenvalue Optimization

Shape and topology optimization using eigenvalue sensitivity analysis in structural mechanics and photonics.

Weak-Strong Form Hybrid Coupling

Coupling finite element and finite difference methods in different subdomains for computational flexibility.

Transformed Field Expansion Methods

Coordinate transformation techniques for efficiently solving problems with localized features or anisotropy.

Isogeometric B-spline Collocation

Direct collocation using CAD-compatible B-spline basis functions for geometry-preserving analysis.

Relaxation Methods for Nonlinear Systems

Fixed-point and nonlinear Jacobi iterations with damping and acceleration strategies for convergence.

Stochastic Galerkin Projection

Intrusive uncertainty quantification via polynomial chaos expansion in the Galerkin weak form.

Entropy Stable Schemes

Discretizations satisfying discrete entropy inequalities for nonlinear conservation laws and hyperbolic systems.

Partition of Unity Localization

Local basis function construction using partition of unity for dimension-independent meshfree approximations.

Finite Difference-Finite Element Hybrid

Coupling finite difference and finite element discretizations in different regions for combined efficiency.

Kinetic Upwind Numerical Methods

Schemes derived from kinetic theory incorporating upwind fluxes for hyperbolic conservation laws.

Certified Reduced Basis Surrogate

Construction of parameterized reduced models with guaranteed error bounds for rapid design evaluation.

Compatible Flux Reconstruction Schemes

Development of high-order shock-capturing schemes combining stability with accuracy preservation.

Multisymplectic Formulation Discretization

Numerical methods preserving multisymplectic structure for Hamiltonian PDEs and wave-like equations.

Adjoint Sensitivity Analysis Methods

Efficient computation of parameter sensitivities using adjoint PDE formulations for inverse problems.

Cut Finite Element Methods

Unfitted discretizations on background meshes with interface handling for moving boundary problems.

Floating-Point Error Analysis

Rigorous treatment of rounding errors and backward stability for numerical algorithm certification.

Certified Reduced Order Models for Real-Time Control

Development of rigorous error bounds and fast construction techniques for reduced basis methods enabling real-time parametric control of high-dimensional dynamical systems.

Locally Enriched Partition Decomposition

Adaptive enrichment of domain decomposition basis functions for improved convergence on heterogeneous media.

Semi-Definite Programming Relaxations

Convex relaxations of nonconvex optimization problems with duality gaps and tightness analysis.

Implicit-Explicit Schemes for Multiphysics Coupling

Design and analysis of decoupled time-stepping strategies that treat stiff and non-stiff components separately while maintaining stability and accuracy in coupled multiphysics problems.

Localized Orthogonal Decomposition for Heterogeneous Media

Construction of multiscale basis functions through local spectral problems to efficiently solve PDEs in materials with highly oscillatory or rough coefficients.

Transient Simulation Error Control

Adaptive time-stepping and mesh refinement with guaranteed bounds on temporal discretization errors.

Localized Non-Conforming Basis

Construction of discontinuous basis functions with optimal local approximation for coarse-grid correction.

Topology Optimization via Level Set Evolution Methods

Integration of level set methods with shape calculus and numerical optimization algorithms for topology optimization problems in engineering design and inverse problems.

Hybrid Finite Difference-Neural Network Solvers

Integration of classical finite difference discretization schemes with trainable neural network components to balance interpretability, efficiency, and accuracy in PDE solving.

Convex Splitting Energy Stable Methods

Splitting of energy functionals into convex and concave parts for provably stable time stepping.