Resource-bounded multi-agent systems in uncertain environments

Introduction

An exciting opportunity to apply for a funded full-time PhD studentship in the Faculty of Environment and Technology, UWE Bristol. The studentship will be funded by the Computer Science Research Centre (CSRC), UWE Bristol. Ref: 2021-OCT-FET08

The expected start date of this studentship is 1 October 2021.

The closing date for applications is Thursday 1 July 2021. 

About the studentship

The Computer Science Research Centre (CSRC) at UWE Bristol is pleased to offer this PhD studentship.

An agent in a multi-agent system (MAS) is considered as an autonomous entity that is capable of reasoning about its behaviour and interaction, and making decisions (to or not to perform actions) on its own in pursuit of its own goals or while acting on behalf of a user or other agents in the system.

Many real-world systems are essentially distributed and complex, such as Internet of Things (IoT) and Cyber Physical Systems (CPS). They are deeply rooted in activities of our daily living and can naturally be modelled as multi-agent systems. These systems are currently becoming increasingly important in the field of robotics, health care systems (in general) and medical rehabilitation (in particular), non-driver transport systems such as UAVs, and in sensor networks, among others.

Many such systems are inherently probabilistic as well as resource-bounded (computationally as well as communication-wise). They usually operate in unpredictable and/or uncertain environments. The uncertainties can make such systems stochastic as well as nondeterministic, depending on the availability of information about the distribution of choices of the actions. The uncertainties can also make such systems adversarial when they involve multiple agents with potentially conflicting goals.

The extremely complex components of such systems and the interaction between them can thus lead to undesired behaviours, and there is no straightforward way of reasoning about their uncertain and unpredictable behaviour. In fact, due to inaccuracy in modelling the real-world probabilistic resource-bounded agents, we cannot say that a system will be absolutely safe and reliable or will definitely be achieving the desired goals.

Aims and objectives

The aim of the proposed project is to define a framework for the representation, specification and verification of probabilistic resource-bounded agents. The objectives of the project are:

  • to define epistemic logics and their computational models for the representation of resource-bounded reasoning agents in a cooperative and probabilistic setting. The resources could be money, energy, or anything else which may be required by an agent for performing an action
  • to investigate the fundamental principles of resource-bounded probabilistic agent verification, and efficient model-checking algorithms for the resulting logics
  • to demonstrate the scalability and expressiveness of our approach by applying the framework and tool to one or more application scenarios.

UWE Bristol

At UWE Bristol, CSRC's research efforts cover many of the facets of 21st century computer science, from AI and IoT device graphene sensors application, to cyber security and autonomous systems. The Centre has more than ten PhD students and a number of RA/KTP associates. Bristol (University of Bristol and UWE Bristol) is home to vibrant theoretical computer science and intelligent systems research communities, and UWE Bristol is an active participant in the UK-RAS NETWORK Robotics and Autonomous Systems. 
 
For an informal discussion about the studentship, please contact Dr Rakib Abdur at Rakib.Abdur@uwe.ac.uk or Professor Jun Hong at Jun.Hong@uwe.ac.uk.

 

 


Funding

The studentship is available from 1 October 2021 for a period of three years, subject to satisfactory progress, and includes a tax-exempt stipend, which is currently £15,609 per annum.

In addition, full-time tuition fees will be covered for up to three years (UK/EU rates only). Overseas applicants are welcome to apply but will be required to cover the difference between Home/EU and overseas tuition fee rates in each year of study.

Eligibility

Applicants must have a good honours degree (2:1 or equivalent) in Computer Science or a closely related discipline, with a research interest in the areas related to logics in computer science, multi-agent systems and formal verification. Applicants must also have a strong background in programming languages, especially in Python and/or Java, together with a basic knowledge of temporal logics and formal verification (model checking and/or theorem proving). Prior knowledge of the PRISM model checker is preferred but not essential. 

A recognised English language qualification is required.

How to apply

Please submit your application online. When prompted use the reference number 2021-OCT-FET08.

Supporting documentation: you will need to upload your research proposal, all your degree certificates and transcripts and your proof of English language proficiency as attachments to your application so please have these available when you complete the application form.

Research proposal: please explain, in at most two A4 pages, how your skill-set and experience could contribute to this proposed PhD project.

References: you will need to provide details of two referees as part of your application. At least one referee must be an academic referee from the institution that conferred your highest degree. Your referee will be asked for a reference at the time you submit your application, so please ensure that your nominated referees are willing and able to provide references before you submit your application. 

Closing date

The closing date for applications is Thursday 1 July 2021.

Further information

Interviews will take place within 4 weeks of the closing date for applications. If you have not heard from us by this date then we thank you for your application but on this occasion you have not been successful.