Current students:
Recent completions:
Overview of Current Students:
Anush Poghosyan, since February 2007, under the supervision of Dr Vadim Zverovich.
The aim of Anush Poghosyan's PhD project is to develop the probabilistic method in order to prove upper bounds for various graph-theoretic parameters. Another aim is to disprove a number of known conjectures using a computer-aided approach.
Tom Richardson, since April 2007, under the supervision of Ana Sendova-Franks.
The aim of Tom Richardson's PhD thesis is to contribute to our understanding of the self-regulatory mechanisms of complex biological systems using experimentation on ant colonies and modelling based on individual threshold distributions for task performance in the context of division of labour. His studentship is part of a multidisciplinary project on how self-regulatory systems work involving physicists (Professor Kim Christensen and Dr Elsa Arcaute, Imperial College), social scientists (Dr Angela Espinosa and Pedro Pablo Cardoso, University of Hull) and roboticists (Dr Torbjorn Dahl and Faruque Sarker, University of Wales at Newport). Tom Richardson is based at the Ant Lab at the University of Bristol and is co-supervised by Professor Nigel Franks.
Tom Richardson is funded through Sendova Frank's recent EPSRC grant "Defying the rules: how self-regulatory social systems work" [Click here for EPSRC external link]
Martin Serpell, since October 2007, under the supervision of Dr Alistair Clark.
The UK’s Office of National Statistics (ONS) often needs to protect the confidentiality of “sensitive” data in published tables, achieving this via a number of approaches collectively known as Statistical Disclosure Control (SDC). The sheer size of the tables means that existing methods are no longer possible to use. This project concerns the development of new mathematical models of how accurately a malicious “attacker” can estimate the value of data in sensitive cells in a table. Building on this the project will then develop techniques to create tables in which confidentiality is protected. The project will build on an existing collaboration between UWE and ONS which has generated experience in this area, and substantial research expertise in Mathematical Modelling (Dr Clark at UWE) and Evolutionary Computation (Dr Smith at UWE).
Zahid Shareef, since February 2007, under the supervision of Dr Rob Laister.
The aim of Zahid Shareef's PhD project is to understand the effects of non-locality (e.g. infection delay or transmission at a distance) dynamics of deterministic mathematical models of infectious diseases. In particular they seek to understand the influence of system parameters (e.g. infection rate, diffusion rate, incubation time) on the qualitative solution structure of the models (e.g. convergence to spatial patterns, periodic outbreaks, epidemic waves).
Yasser Shehata, since 2004, under the supervision of Dr Paul White.
Yasser Shehata is a part-time PhD student studying with Paul White and Alistair Clark. Yasser's PhD programme of study concerns the correction of problems associated with the use of automated computer in the development of regression modelling solutions. Recent peer reviewed conference papers include:
Shehata, Y A and White, P (2007), A randomization method to correctly estimate overall significance in best subsets regression, 17th International Conference on Computer Theory and Applications, Alexandria.
Shehata, Y A and White, P (2007) A computational approach to correctly assess significance in best subset regression, International Arab Conference on Information Technology 2007, Syria.
Yasser is a lecturer in statistics at the Arab Academy for Science, Technology, and Maritime Transport Productivity and Quality Institute in Alexandria and takes a lead role in the teaching of Six Sigma and Quality Assurance.Andrea Staggemeier, since 2001, under the supervision of Dr Alistair Clark.
The aim of the thesis was to better understand the use of hybrid approaches (i.e. combination of mathematical programming techniques and metaheuristic techniques) to solve complex problems such as the single-stage multi-machine simultaneous lot sizing and scheduling problem . The research demonstrates a successful implementation of hybrid evolutionary algorithm at little expense of either loss of generality, or inability to solve large instances.
Eli Toso, PhD student at the Federal University of São Carlos, Brazil and visiting student at UWE, Jan–May 2006, under the co-supervision of Dr Alistair Clark. Soon to defend.
The thesis proposes a mixed integer programming model for joint lot sizing and scheduling to a plant for animal feed compounds. A key characteristic of this industry is that certain products can perform a production line cleaning function if a sufficiently large lot is produced between two products that would otherwise require a cleaning setup. Thus the sequence-dependent setup times do not always obey the triangular inequality. Tested on data from the plant, the model takes too long to solve exactly and so several alternative formulations and methods are developed to solve the model more quickly, based on variants of the Relax and Fix heuristic and the Asymmetric Travelling Salesman Problem (ATSP). The results confirm that the formulations are computationally effective and able to take economic advantage of the intermediate cleaning products. The models' schedules substantially improves on those practiced at the feed plant.
Overview of Recent Completions:
Modelos e Métodos Da Programação da Produção Aplicados no Setor de Fundições
(Models and Methods for Production Scheduling in Foundries).
Defended May 2003. Co-Supervisor: Dr Alistair Clark.
PhD student at the University of São Paulo, Brazil, visiting CEMS, UWE, March – July 2001.
This thesis researched a lot sizing and scheduling problem from a foundry n which key materials are produced and then transformed into many products on a single machine. A mixed integer programming (MIP) model was developed, taking into account sequence-dependent setup costs and times, and then adapted for rolling horizon use. A relax-and-fix (RF) solution heuristic was proposed and computationally tested against a high-performance MIP solver. Three variants of local search were also developed to improve the RF method and tested. Finally the solutions were compared with those currently practiced at the foundry.
Stability of Interfacial Flows.
Defended July 2006 Supervisor: Dr Alison Hooper.
Satish's thesis investigated the flow instability in terms of disturbance energy growth following techniques developed by Henningson et al. He extended the two-dimensional work of South and Hooper of two superposed viscous fluids to take into account three-dimensional effects. Problems of non-convergence, previously identified by South and Hooper as being caused by the interfacial eigenmode, are overcome by replacing the interface by a miscible layer. The work within the thesis involved extensive use of the numerical collocation technique applied to the governing three-dimensional stability equations of two and three layer vsicous fluid flow.
Multi-objective approach for university staff planning (an empirical study of an Egyptian private university).
Defended December 2005. Supervisor: Dr Alistair Clark.
The thesis studied multi-criteria decision-making models for manpower planning of teaching staff planning in Egyptian Higher Education. The research developed a goal-programming model to let managers explore the trade-offs between targets associated with economic efficiency and academic quality while at the same time taking into account constraints such as the availability of scarce resources and penalising the amount by which targets might be missed. Survey results from a particular university were used in the formulation process and to evaluate the usefulness of the model results. The thesis extended several multi-objective mathematical optimisation approaches for the difficult task of mapping the so-called “efficientfrontier”, i.e., the multidimensional surface of optimal trade-offs which decision-makers need to identify in order to develop their understanding of complex trade-off relationships. The result was a powerful tool to help university management make better choices.
Weight approximations in multi-attribute decision models.
Defended September 2004. Supervisor: Supervisor: Dr Alistair Clark.
The thesis concerned multi-criteria decision making, something we all do when making decisions such as which house to buy, or where to go on holiday. Such decisions inevitably involve tradeoffs between conflicting criteria (or attributes). A simple yet powerful method that is often used is the Simple Multi-attribute Rating Technique using Swings (SMARTS). Ron investigated the issue of how criteria should be weighted when summing the criterion scores of the different decision options with SMARTS. His original research contribution was to offer a mathematical explanation for the consistent differences in the distribution of weights when certain alternative weighting methods are used. In particular, Ron theoretically demonstrated the advantages of Rank Order Distribution (ROD) weights over the conventional alternative Rank Order Centroid weights. In the process he discovered the formulae for a family of piecewise probability density functions representing the distributions of ranked weights. The final part of his PhD research backed up his theoretical analysis with experiments concerning how people actually assess weights in practice.
Brood sorting in the ant Temnothorax alpipennis: from biology to collective robotics.
Defended 2005. Supervisor: Dr Ana Sendova-Franks.
Samuel Scholes' thesis aimed to understand the mechanisms underlying brood sorting in ants with the view of informing sorting algorithms for virtual and embodied groups of autonomous robots. Rock ants sort their brood in concentric annuli of different developmental stages so that those items that need most care are on the periphery. Sam’s work combined novel manipulative experiments with ant colonies, such as testing whether gaseous templates are a necessary condition for generating the sorted pattern, with experiments on embodied robots to test the hypothesised algorithm of individual behaviour followed by ants. The publications from his thesis demonstrate that brood sorting in ants has two phases: clustering followed by spreading, a type of a non-linear diffusion process. These results have had impact on the robotics community and have inspired further theoretical studies and modelling by mathematicians and physicists. Sam was co-supervised by Professor Chris Melhuish, Director of the Bristol Robotics Laboratory (UWE and University of Bristol).'
Any comments to this web page should be directed to the AMG Webmaster