Reinforcement Learning (RL) relies on a reward function as the only feedback in an interaction to learn complex tasks. Multi-objective RL (MORL) uses multiple reward functions. The aim of this thesis will be to produce new MORL algorithms that are capable to produce optimal and ethically aligned actions.
PhD Projects SEIT
Scholarships of $35,000 (AUD) are available for PhD students who have achieved Honours 1/High Distinction in their UG program and/or have completed a Masters by Research.
Defence systems acquisition is fraught with all sorts of financial, technical and political risks. The most effective way of mitigating risks associated with acquisition of complex systems is through identification of these risks as early as possible. This project’s aim is to facilitate optimal decision making for multi stakeholder system design.
Renewable energy has shown promising future of practical applications. Advanced learning and control technology will enable efficient renewable energy systems. This project will explore learning control technologies to enhance the efficiency in renewable energy systems.
Accelerating hypersonic flight is characterised by rapidly changing flow conditions and high heat loads. To achieve sustained combustion in these engines, fuel injectors must be capable of also rapidly responding to changing flow conditions to ensure efficient and effective fuel delivery is maintained.
With the spread of IoT devices, security issues are becoming more severe, in part because of the large scale and heterogeneous nature of the devices.
There are an increasing number of insecure IoT devices with a high computational power, this makes them attractive targets for botnet creators.
Compromised IoT devices can be aggregated together through command and control servers to perform a diverse set of activities including; distributed denial of service, password cracking, and crypto-currency mining.
Achieving appreciable thrust for both cruise and accelerating flight in scramjet engines is only half the solution for a viable engine. Equally important, is the design and validation of a vehicle structure capable of withstanding the sustained high thermal loads, and the resulting thermally induced structural stresses.
Over the years, we have developed state-of-the-art distributed multi-agent system capable of modelling reciprocal competitive interactions between blue and red teams. Our systems can reason in real-time and explain emerging behaviours. This project will design distributed artificial intelligence algorithms for the agents within our system.
Learning and teaching fully online is a challenging task as it requires maximum engagement with fellow students and the course convenor to create a sense of online community among distance learners.
In contemporary enterprises, single project settings are rare today. Hence, issues involving the simultaneous management of multiple projects (or portfolio of projects) have become more prevalent. Up to 90 % of all projects worldwide are executed in a multi-project context (i.e., portfolio of projects). This portfolio of projects (POP) is considered as the simultaneous scheduling of two or more projects which demand the same scarce resources.
Metasurfaces are a promising new area of science and technology for the manipulation of electromagnetic waves. They have the potential to replace bulky devices such as lenses with a thin layer of patterned elements, to enable a wave to undergo focussing, steering, polarisation conversation and more general spatial/temporal filtering.