Skip to main content

Research Repository

Advanced Search

Supervisions (9)

PhD
Doctorate

Level Doctorate
Student Katarina Alexander
Status Withdrawn
Part Time Yes
Years 2021
Project Title Unifying Data Driven Decision Making and Evolutionary Computing
Awarding Institution Edinburgh Napier University
Director of Studies Shufan Yang
Second Supervisor Emma Hart

PhD
Doctorate

Level Doctorate
Student Mr Magnus Janson
Status Current
Part Time No
Years 2021
Project Title Assessing biodiversity and ecological status of European flat oyster beds using soundscape analysis, eDNA and visual surveys
Awarding Institution Edinburgh Napier University
Director of Studies Karen Diele
Second Supervisor Emma Hart
Additional Supervisor Jennifer Dodd

PhD
Doctorate

Level Doctorate
Student Richard Plant
Status Current
Part Time No
Years 2020
Project Title Privacy-preserving methods for Natural Language Processing
Awarding Institution Edinburgh Napier University
Director of Studies Dimitra Gkatzia
Additional Supervisor Emma Hart

PhD
Doctorate

Level Doctorate
Student Kirsty Montague
Status Current
Part Time Yes
Years 2019
Project Title Evolving Robust Behaviours for Robotic Swarms with a Modular Design Approach
Awarding Institution Edinburgh Napier University
Director of Studies Emma Hart
Second Supervisor Ben Paechter
Additional Supervisor Andreas Steyven

PhD
Doctorate

Level Doctorate
Student Grant Anderson
Status Current
Part Time Yes
Years 2019
Project Title Augmenting Neural Attention Models in Conversation Modelling
Awarding Institution Edinburgh Napier University
Director of Studies Emma Hart
Second Supervisor Dimitra Gkatzia

PhD
Doctorate

Level Doctorate
Student Dr Kehinde Babaagba
Status Complete
Part Time No
Years 2017 - 2021
Project Title Application of evolutionary machine learning in metamorphic malware analysis and detection
Awarding Institution Edinburgh Napier University
Director of Studies Thomas Tan
Second Supervisor Emma Hart

PhD
Doctorate

Level Doctorate
Student Dr Andreas Steyven
Status Complete
Part Time No
Years 2013 - 2018
Project Title A closer look at adaptation mechanisms in simulated environment-driven evolutionary swarm robotics
Project Description Swarm robotics is a special case within the general field of robotics. The distributed nature makes it more resilient with no single point of failure. Considering the application in remote locations, the swarm needs to adapt autonomously to a priori unknown environmental conditions. A special branch of evolutionary robotics achieves online adaptation during runtime through embedding the evolutionary algorithm in the robot. A well studied algorithm to do that is mEDEA, minimal Environment-driven Distributed Evolutionary Adaptation, which has been shown to be able to maintain the swarm's integrity in an abruptly changing environment. It’s completely decentralised nature and the lack of an explicit fitness function leaves only the environment to provide the driving force for the evolutionary process.

This thesis investigates several aspects of environment-driven adaptation in simulated evolutionary swarm robotics. It is centred around a specific algorithm for distributed embodied evolution called mEDEA.
Firstly, mEDEA is extended with an explicit relative fitness measure while still maintaining the distributed nature of the algorithm. Two ways of using the relative fitness are investigated: influencing the spreading of genomes and performing an explicit genome selection. Both methods lead to an improvement in the swarm’s ability to maintain energy over longer periods.
Secondly, a communication energy model is derived and introduced into the simulator to investigate the influence of accounting for the costs of communication in the distributed evolutionary algorithm where communication is a key component.
Thirdly, a method is introduced that relates environmental conditions to a measure of the swarm’s behaviour in a 3-dimensional map to study the environment’s influence on the emergence of behaviours at the individual and swarm level. Interesting regions for further experimentation are identified in which algorithm specific characteristics
show effect and can be explored.
Finally, a novel individual learning method is developed and used to investigate how the most effective balance between evolutionary and lifetime-adaptation mechanisms is influenced by aspects of the environment a swarm operates in. The results show a clear link between the effectiveness of different adaptation mechanisms and environmental conditions, specifically the rate of change and the availability of learning opportunities.
Awarding Institution Edinburgh Napier University
Director of Studies Emma Hart
Second Supervisor Ben Paechter
Thesis A Closer Look at Adaptation Mechanisms in Simulated Environment-Driven Evolutionary Swarm Robotics

PhD
Doctorate

Level Doctorate
Student Dr Kevin Sim
Status Complete
Part Time Yes
Years 2010 - 2014
Project Title Novel hyperheuristics applied to the domain of bin packing
Project Description Hyper-heuristics (HH) have been described as methodologies that aim to offer “good enough -soon enough - cheap enough” solutions to real world problems.

Many real world problems can be formulated as combinatorial optimisation problems in which the permutation of the problem elements defines the quality of the solution. Common examples include routing, scheduling, timetabling, packing and constraint satisfaction problems which are often combined in real world situations, making the problems more difficult and the methods employed to solve them less robust and less transferable. Problem specific, bespoke applications often prove too costly and time consuming to implement for many real world business applications with little scope for reuse.

The primary goal of the research being conducted is to investigate novel hyper-heuristic approaches for solving combinatorial optimisation problems with the goal of developing novel classification approaches that map a problems composition to the suitability of simple domain specific heuristic techniques for solving the problem.

One recent grouping of HH approaches is to separate them into two categories described as “Heuristics to Select Heuristics" and “Heuristics to Generate Heuristics"

To date success has been gained while investigating the first category using classification algorithms to select, based on a problem instances characteristics, which from a set of domain specific heuristics will fare best.

Current research is focussing upon generating heuristics using evolutionary programming techniques to incorporate into the system developed to date with the eventual aim of combining automated heuristic design and heuristic selection techniques into a continually adapting problem solver.
Awarding Institution Edinburgh Napier University
Director of Studies Emma Hart
Second Supervisor Ben Paechter
Thesis Novel Hyper-heuristics Applied to the Domain of Bin Packing