Human-Centric Optimization: A Data Enabled Decision Making Perspective - Computer Science - EPSRC DTP funded PhD Studentship

Scheduling, vehicle routing, network configuration, and resource allocation are examples of hard optimization problems with broad application in industry. The focus of the past research in this area has almost been on algorithmic issues. However, this approach neglects many important human-computer interaction issues that must be addressed to provide practitioners with engineering-intuitive, practical solutions to optimisation problems. Automatic methods do not leverage human expertise and can only find solutions that are optimal with regard to an invariably over-simplified problem description. Furthermore, users must understand the generated solutions in order to implement, justify, or modify them. Interactive optimization helps address these issues but has not previously been studied in detail. Bearing this consideration in mind, this project will develop automatic methods to find solutions that are not only optimal, but also preferred by the decision maker (DM). In particular, this will be implemented through “human in the loop” interactive optimization. Specifically, it has the following two inter-linked work packages.

Due to the lack of knowledge of the underlying optimization problem, it is inevitable to incorporate uncertainty in this interactive optimization process. The first step of this project is to quantitatively measure the DM’s vagueness of the information or imprecise judgements. Gaussian process (GP), a viable tool to handle noise and uncertainty, will be used to serve this purpose. In order to alleviate overfitting, we need to jointly estimate both the GP model discrepancy and the “best” input values.

To develop an interactive optimization framework, which is able to lead a DM to her/his solution(s) of interest (SOI). The progress towards the most preferred solution is made by accepting preference-based information progressively from the DM after every few iterations of an algorithm. In particular, it is essential to investigate what does a good decision look like and how to measure it? Afterwards, this preference information is used to model a strictly monotone value function, which is used for the subsequent iterations of the optimization. Note that due to the uncertainty and noise in the decision-making process, especially when a DM is in presence of imperfect data (i.e., not fully converged population), we will apply the uncertainty model developed in the first WP to aggregate the value function. Thereafter, the constructed value function can be readily used for fitness assignment and thus direct an algorithm’s search towards SOI.

Candidate

The project would suit a student with a first degree in Computer Science and has a desire to develop a range of skills such as evolutionary multi-objective optimisation, human-computer interaction, machine learning and data mining. Candidates with the background of evolutionary computation, machine learning and data mining are highly encouraged to apply.

Entry Requirements:

You should have or expect to achieve at least a 2:1 Honours degree, or equivalent, in [computer science or engineering]. Experience in [evolutionary computation, human-computer interaction, visualization, machine learning and data mining, computer networks] 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.