Research Interests

I am a numerical analyst with a focus on surrogate modelling in the context of uncertainty quantification. My research has focused on non-intrusive techniques for global polynomial approximations. In particular, I am interested developing efficient adaptive methodologies that can balance the computational cost and approximation errors within the different approximation stages. My research spans simple analytic problems with provable properties to challenging computational fluid dynamics models.

Highlights and available software include:

• Development of novel multi-fidelity stochastic collocation algorithm with a focus on “noisy” solvers. This is part of the Sparse Grids MATLAB Kit project.
• Hierarchical based error estimation for sparse grid polynomial approximation of parametric parabolic PDE available at https://github.com/benmkent/adaptive_sc_fem. Development and implementation of novel residual based strategy is ongoing.
• Containerisation of models for Non-Intrusive Uncertainty Quantification. Contributed a benchmark problem to the UM-BRIDGE project. Further research models are available at https://github.com/benmkent/umbridge-servers.
• Implementation of sparse grid interpolation and quadrature techniques in Julia. This is available at https://github.com/benmkent/SparseGridsKit.jl.