The numerical approximation of parametric partial differential equations is acomputational challenge, in particular when the number of involved parameter is large.This paper considers a model class of second order, linear, parametric, elliptic PDEs on abounded domain D with diffusion coefficients depending on the parametersin an affine manner. For such models, it was shown in [9, 10] that under very weak assumptionson the diffusion coefficients, the entire family of solutions to such equations can besimultaneously approximated in the Hilbert space V = H01(D) by multivariate sparse polynomials in the parametervector y with a controlled number N of terms. Theconvergence rate in terms of N does not depend on the number ofparameters in V, which may be arbitrarily large or countably infinite,thereby breaking the curse of dimensionality. However, these approximation results do notdescribe the concrete construction of these polynomial expansions, and should thereforerather be viewed as benchmark for the convergence analysis of numerical methods. Thepresent paper presents an adaptive numerical algorithm for constructing a sequence ofsparse polynomials that is proved to converge toward the solution with the optimalbenchmark rate. Numerical experiments are presented in large parameter dimension, whichconfirm the effectiveness of the adaptive approach.

We consider a two-dimensional quantum waveguide composed of two semi-strips of width 1and 1 − ε, where ε > 0 is a small real parameter,i.e. the waveguide is gently converging. The width of the junction zonefor the semi-strips is 1 + O(√ε). We will present a sufficient condition for the existence of a weaklycoupled bound state below π2, the lower bound of thecontinuous spectrum. This eigenvalue in the discrete spectrum is unique and itsasymptotics is constructed and justified whenε → 0+.

Space-time approximations of the FitzHugh–Nagumo system of coupled semi-linear parabolicPDEs are examined. The schemes under consideration are discontinuous in time butconforming in space and of arbitrary order. Stability estimates are presented in thenatural energy norms and at arbitrary times, under minimal regularity assumptions.Space-time error estimates of arbitrary order are derived, provided that the naturalparabolic regularity is present. Various physical parameters appearing in the model aretracked and numerical examples are presented.

We propose a new reduced basis element-cum-component mode synthesis approach forparametrized elliptic coercive partial differential equations. In the Offline stage weconstruct a Library of interoperable parametrized reference componentsrelevant to some family of problems; in the Online stage we instantiate andconnect reference components (at ports) to rapidly form and query parametricsystems. The method is based on static condensation at the interdomainlevel, a conforming eigenfunction “port” representation at the interface level, andfinally Reduced Basis (RB) approximation of Finite Element (FE) bubble functions at theintradomain level. We show under suitable hypotheses that the RB Schur complement is closeto the FE Schur complement: we can thus demonstrate the stability of the discreteequations; furthermore, we can develop inexpensive and rigorous (system-level) aposteriori error bounds. We present numerical results for model many-parameterheat transfer and elasticity problems with particular emphasis on the Online stage; wediscuss flexibility, accuracy, computational performance, and also the effectivity of thea posteriori error bounds.

This paper presents a numerical investigation of plaque growth in a diseased artery using the two-way fluid–structural interaction (FSI) technique. An axis-asymmetric 45% stenosis model is used as the base model to start the plaque growth approximation. The blood is modelled as a non-Newtonian fluid described by the Casson model. The artery tissue is assumed to be a nonlinear material. The two-way FSI simulation is carried out in a way that mimics the unsteady blood flow through a diseased artery by using a pulsatile flow condition. After each flow velocity cycle, the numerical results are extracted and used to modify the stenosis geometry based upon critical wall shear stress (WSS) values and an accepted relationship between the concentration of low density lipoprotein and WSS. The simulation procedure is repeated until the growth-updated stenosis morphology reaches 79% severity. The behaviour of the flow velocity is analysed at each growth stage, together with the WSS, to determine the change of plaque morphology due to growth. The effects of WSS and pressure on the artery wall at the final stage (79% severity) of the plaque growth model are also compared with results from the authors’ previous work, to demonstrate the importance of the morphology change in plaque growth modelling.

The vertical rise of a round plume of light fluid through a surrounding heavier fluid is considered. An inviscid model is analysed in which the boundary of the plume is taken to be a sharp interface. An efficient spectral method is used to solve this nonlinear free-boundary problem, and shows that the plume narrows as it rises. A generalized condition is also introduced at the boundary, and allows the ambient fluid to be entrained into the rising plume. In this case, the fluid plume first narrows then widens as it rises. These features are confirmed by an asymptotic analysis. A viscous model of the same situation is also proposed, based on a Boussinesq approximation. It qualitatively confirms the widening of the plume due to entrainment of the ambient fluid, but also shows the presence of vortex rings around the interface of the rising plume.

This is a review of thin-body and slender-body theories, with indications of some new applications. Topics discussed include bodies with near-constant surface pressure, subsonic and supersonic aerodynamics, ship hydrodynamics, slender bodies in Stokes flow, slender footings in elastic media, and slender moonpools. Mathematical features of the thin- and slender-body approximations are also discussed, especially nonlocal convolution terms modelling three-dimensionality in the otherwise two-dimensional near field, end effects, and the role of the logarithm of the slenderness ratio. This review was presented by the first author as the IMA Lighthill Memorial Lecture at the British Applied Mathematics Colloquium (BAMC) 2004.

Infecting Aedes aegypti mosquitoes with the bacteria Wolbachia has been proposed as an innovative new strategy to reduce the transmission of dengue fever. Field trials are currently being undertaken in Queensland, Australia. However, few mathematical models have been developed to consider the persistence of Wolbachia-infected mosquitoes in the wild. This paper develops a mathematical model to determine the persistence of Wolbachia-infected mosquitoes by considering the competition between Wolbachia-infected and non-Wolbachia mosquitoes. The model has four steady states that are biologically feasible: all mosquitoes dying out, only non-Wolbachia mosquitoes surviving, and two steady states where non-Wolbachia and Wolbachia-infected mosquitoes coexist. The stability of the steady states is determined with respect to the key parameters in the mosquito life cycle. A global sensitivity analysis of the model is also conducted. The results show that the persistence of Wolbachia-infected mosquitoes is dominated by the reproductive rate, death rate, maturation rate and maternal transmission. For the parameter values where Wolbachia persists, it dominates the population, and hence the introduction of Wolbachia has great potential to reduce dengue transmission.