# SIAM UKIE National Student Chapter Conference

## 18-19 June 2018

The University of Bath SIAM-IMA Student Chapter is privileged to be hosting the SIAM UKIE National Student Chapter Conference in 2018. The conference will bring together students working in all areas of applied and industrial mathematics, and related fields, from across the UK and Ireland.

NEWS: Deadline for contibuted talks and posters extended to Friday 27th April.

NEWS: Funding guaranteed for student speakers. In some cases we may also be able to reimburse poster presenters.

# Submissions

Deadline for contibuted talks and posters extended to Friday 27th April.

We are now accepting submissions for contributed talks and posters. Please submit your title and abstract via the registration form.

Funding is available to support students giving presentations. If you would like to apply for this funding, please note this in the registration form. We aim to refund travel and accommodation for all speakers, please indicate how much you expect your travel to cost. In some cases we may be able to reimburse poster presenters, please also indicate if you require funding.

The deadline for contributed talk abstracts is Friday 30th March 2018 Friday 27th April, and for poster submissions Friday 13th April 2018 Friday 27th April.

We follow both the SIAM Statement on Inclusiveness and the IMA Diversity Statements, and we endeavour to create a welcoming and inclusive atmosphere at the conference, with a diverse range of participants. Students from groups which are underrepresented in the mathematical sciences are particularly encouraged to submit an abstract for a talk or poster.

# Registration

Registration is now open. Please complete the registration form by Friday 18th May 2018.

As noted above, funding is available to support students giving presentations. If you would like to apply for this funding, please note this in the registration form and we will do our best to accommodate your needs.

If you have any questions about the conference, you're welcome to email the Bath SIAM-IMA Student Chapter Committee.

# Programme

The programme will comprise sessions of student talks and posters, and four plenary talks from the folowing speakers:

 Monday 18th June Tuesday 19th June 08:30 -- 09:00 COFFEE 09:00 -- 09:30 SESSION 3AStudent TalksWolfson Lecture Theatre SESSION 3BStudent Talks4 West 1.2 09:30 -- 10:00 10:00 -- 10:30 COFFEE & REGISTRATION COFFEE 10:30 -- 11:00 WelcomePlenary Talk - Nicole SpillaneDomain Decomposition Methods with Adaptive MultipreconditioningUniversity Hall Plenary Talk - Ian GriffithsiPhones and Dysons: using fluid dynamics to tailor technologyUniversity Hall 11:00 -- 11:30 11:30 -- 12:00 SESSION 4AStudent TalksWolfson Lecture Theatre SESSION 4BStudent Talks4 West 1.2 12:00 -- 12:30 SESSION 1AStudent TalksWolfson Lecture Theatre SESSION 1BStudent Talks4 West 1.2 12:30 -- 13:00 LUNCH4 West Atrium 13:00 -- 13:30 LUNCH4 West Atrium 13:30 -- 14:00 SESSION 5AStudent TalksWolfson Lecture Theatre SESSION 5BStudent Talks4 West 1.2 14:00 -- 14:30 SESSION 2AStudent TalksWolfson Lecture Theatre SESSION 2BStudent Talks4 West 1.2 14:30 -- 15:00 Plenary Talk - Simon Chandler-WildeCoercivity of second kind boundary integral equations on Lipschitz domainsUniversity HallClose 15:00 -- 15:30 COFFEE 15:30 -- 16:00 Plenary Talk - Marta BlangiardoModelling spatio-temporal data: methods, examples and challengesUniversity Hall 16:00 -- 16:30 16:30 -- 17:00 Poster Session&Wine Reception4 West Atrium 17:00 -- 17:30 17:30 -- 18:00 19:00 DINNER - Bath City Centre

In addition to the conference dinner on the Monday evening, we will organise social events on both the Sunday and Tuesday, for any participants who wish to stay in Bath for an extra evening:

• Sunday evening: Drinks and games in a pub in the city centre (location TBC) to welcome participants arriving on Sunday.
• Tuesday afternoon: Walking tour of the city of Bath, setting off from the university campus at approximately 16:00.

# Abstracts

## Plenary Speakers

### Marta Blangiardo (Imperial College London)

Modelling spatio-temporal data: methods, examples and challenges

In this talk I will present how the Bayesian hierarchical framework is commonly used to model spatial and spatio-temporal data. I will first consider the space as discrete (small area framework) and introduce autoregressive structures, which can be used to account for spatial and temporal dependency. I will focus on different examples drawn from the epidemiological field.

Then I will move to the case when space is continuous and introduce the geostatistics framework. Here I will present two examples drawn from epidemiology and environmental science.

I will finish the talk presenting some of the (methodological and non) challenges which researchers face in this context.

### Simon Chandler-Wilde (University of Reading)

Coercivity of second kind boundary integral equations on Lipschitz domains

Boundary integral equation methods are a well-established technique for solving linear elliptic partial differential equations. Second kind integral equation formulations, taking the form $$\phi + K\phi = g$$, where $$K$$ is a boundary integral operator, are attractive because they are well-conditioned. However, it is not known whether standard Galerkin numerical methods are stable and convergent for these formulations for Laplace's equation with Dirichlet boundary conditions on general Lipschitz domains. When the domain $$D$$ is bounded, smooth, and convex the standard second kind integral equation formulation for this simplest case can be written in this form where $$K$$ is a contraction, i.e. $$\|K\|<1$$. Under certain other conditions on $$D$$ it is known that $$K=L+C$$, with $$\|L\|<1$$ and $$C$$ compact, but it is a long-standing conjecture that this holds for all Lipschitz $$D$$. Both these conditions guarantee convergence of Galerkin approximation methods.

But in this talk we exhibit examples of Lipschitz domains in 2D for which it does not hold that $$K$$ is a compact perturbation of a contraction; in fact the so-called essential norm may be arbitrarily large. On the other hand, in a positive direction, we also exhibit new, modified versions of the standard second kind integral equation formulation which have the attractive property that the associated operators are coercive for all Lipschitz $$D$$, so that every Galerkin method is convergent.

### Ian M Griffiths (Mathematical Institute, University of Oxford)

iPhones and Dysons: using fluid dynamics to tailor technology

As technology continues to advance, new strategies involving a range of scientific disciplines are required. Mathematicians can provide frameworks to predict operating regimes and manufacture techniques. In this talk we present two case studies: the fabrication of precision glass, for smartphones and new flexible devices; and the development of superior filters for vacuum cleaners. In each case we use asymptotic analysis to derive a model that determines the fabrication protocol required to produce a desired final product.

### Nicole Spillane (CNRS, CMAP, École Polytechnique)

Domain Decomposition Methods with Adaptive Multipreconditioning

Domain decomposition methods are a family of parallel solvers for large linear systems. They all share the idea of approximating the inverse of some matrix by a sum of local inverses (in the so-called subdomains). I will present some classical domain decomposition methods, their limitations and some recent efforts to improve their robustness and scalability so that they can be applied to problems arising from real life simulations.

More precisely, I will present the method of adaptive multipreconditioning. This is a modification of the iterative solver (the preconditioned conjugate gradient algorithm). Instead of one single preconditioner, a family of preconditioners is applied at each iteration, each corresponding to one of the subdomains. This significantly increases the size of the minimization space and consequently accelerates convergence. I will introduce the method, discuss its analysis and show numerical results obtained in collaboration with C. Bovet, P. Gosselet and A. Parret-Fréaud.

## Student Talks

A full list of confirmed speakers and abstracts will appear here in Spring 2018.

## Posters

This conference is kindly supported by grants from:

We are also grateful for support from our industrial sponsors:

# Venue

The conference will be held in the Department of Mathematical Sciences (marked as 4W on this campus map).

Department of Mathematical Sciences
4 West
University of Bath
Claverton Down
Bath
BA2 7AY

See the university's guide on how to find us.

## Arriving by train

If you are arriving by train, you should arrive at Bath Spa train stations. You can then take the U1 bus to the university from Dorchester Street (opposite the station to the left), or walk to the university in 30-40 minutes (up quite a steep hill!).

## Arriving into Bristol Airport

Coming from Bristol airport, take the A4 bus to Bath, which brings you to the train station, from where you can walk or take a bus, as described above.

## Arriving by car

If you plan to drive to the university, limited visitor parking spaces are available on campus. See the university's guide to parking on campus.