Automating Knowledge Discovery of Innovative Design Principles - Computer Science - EPSRC DTP funded PhD Studentship.

Multi-objective optimisation problem, which involves optimising more than one conflicting objective, is ubiquitous across the breadth of science, engineering and economics. For example, the cost of fabricating a product and its quality are two typical conflicting objectives of design. The quest for new knowledge about such problems has always been important to engineers and scientists to gain better understandings of the underlying problems. The type and extent of knowledge can be different in different problems, but practitioners are particularly interested in knowing what design principles a solution must have in order for it be an optimal or high-performing solution. Such questions are vitally important to a designer as the answers to such question provide deep insights among parameter interactions that would elevate a design to become optimal.

Project Aims and Methods

The overarching objective of this project is to develop intelligent methods that are able to automate knowledge discovery over high-performing solutions and thus unveil useful design principles to the engineers and practitioners. To this end, this project involves two interlinked steps:

The first step is going to design novel evolutionary multi-objective optimisation (EMO) algorithms to find a set of high-performing trade-off optimal solutions involving two or more conflicting objectives related to the problem. In particular, novel algorithms will be built upon the state-of-the-art decomposition-based EMO framework developed in the supervisors’ groups to handle complex multi- and many-objective optimisation problems. Advanced local search techniques will be developed to further improve the solutions found by the EMO algorithms.

The second step is to apply advanced machine learning and data mining techniques to analyse and understand the common principles hidden in the high-performing solutions found in Step 1. Note that these principles can provide interesting and important design patterns that reveal the relationships among decision variables, objective functions and constraints. In particular, this rule mining process will be formulated as another optimisation problem, either with single objective or multiple objectives depending on the requirements of the decision maker. Advanced machine learning techniques, such as Gaussian processes, will be used to build the model and mitigate the effects of noise and uncertainty within the imperfect training data.

Candidate

This project would suit a student with a first degree in Computer Science, Mathematics or a related subject with a desire to develop a range of skills such as evolutionary multi-objective optimisation, machine learning and data mining. Candidates with the background of evolutionary computation, machine learning and data mining are highly encouraged to apply.

Please include keywords and phrases for which prospective applicants may be searching. We know that funding is the main barrier to study which is reflected in the search volumes, so including additional references to ‘PhD funding’; ‘PhD scholarship’; ‘PhD studentship’ would be of benefit, as well as any other broader potential related search terms.

Entry Requirements
You should have or expect to achieve at least a 2:1 Honours degree, or equivalent, in [computer science, mathematics, engineering or a related subject]. Experience in [evolutionary computation, machine learning and data mining] is desirable.

If English is not your first language you will need to meet the English language requirements and provide proof of proficiency.

The majority of the studentships are available for applicants who are ordinarily resident in the UK and are classed as UK/EU for tuition fee purposes. If you have not resided in the UK for at least 3 years prior to the start of the studentship, you are not eligible for a maintenance allowance so you would need an alternative source of funding for living costs. To be eligible for fees-only funding you must be ordinarily resident in a member state of the EU.

Applicants who are classed as International for tuition fee purposes are NOT eligible for funding. International students interested in studying at the University of Exeter should search our funding database for alternative options.