Transfer Optimisation System for Adaptive Automated Nature-Inspired Optimisation

Hard optimisation problems are ubiquitous across the breadth of science, engineering and economics. For example, in water system planning and management, water companies are often interested in optimising several system performance measures of their infrastructures in order to provide sustainable and resilient water/wastewater services that are able to cope with and recover from disruption, as well as wider challenges brought by climate change and population increase. As a classic discipline, significant advances in both theory and algorithms have been achieved in optimisation. However, almost all traditional optimisation solvers, ranging from classic methods to nature-inspired computational intelligence techniques, ignore some important facts: (i) real-world optimisation problems seldom exist in isolation; and (ii) artificial systems are designed to tackle a large number of problems over their lifetime, many of which are repetitive or inherently related. Instead, optimisation is run as a 'one-off' process, i.e. it is started from scratch by assuming zero prior knowledge each time. Therefore, knowledge/experience from solving different (but possibly related) optimisation exercises (either previously completed or currently underway), which can be useful for enhancing the target optimisation task at hand, will be wasted. Although the Bayesian optimisation considers incorporating some decision maker's knowledge as a prior, the gathered experience during the optimisation process is discarded afterwards. In this case, we cannot expect any automatic growth of their capability with experience. This practice is counter-intuitive from the cognitive perspective where humans routinely grow from a novice to domain experts by gradually accumulating problem-solving experience and making use of existing knowledge to tackle new unseen tasks. In machine learning, leveraging knowledge gained from related source tasks to improve the learning of the new task is known as transfer learning, an emerging field that considerable success has been witnessed in a wide range of application domains. There have been some attempts on applying transfer learning in evolutionary computation, but they do not consider the optimisation as a closed-loop system. Moreover, the recurrent patterns within problem-solving exercises have been discarded after optimisation, thus experience cannot be accumulated over time.

The proposed research will develop a revolutionary general-purpose optimiser (as known as transfer optimisation system) that will be able to learn knowledge/experience from previous optimisation process and then autonomously and selectively transfer such knowledge to new unseen optimisation tasks. The transfer optimisation system places adaptive automation at the heart of the development process and explores novel synergies at the crossroads of several disciplines including nature-inspired computation, machine learning, human-computer interaction and high-performance parallel computing. The outputs will bring automation in industry, including an optimised/shortened production cycle, reduced resource consumption and more balanced and innovative products, which have great potentials to result in economic savings and an increase of turnover. The proposed methods will be rigorously evaluated by the industrial partners, first in water industry and will be expanded to a boarder range of sectors which put the optimisation at the heart of their regular production/management process (e.g. renewable energy, healthcare, automotive, appliance and medicine manufacturers).

This programme of research will create a revolutionary computational search framework. It opens a systematic and rational avenue to the design and analysis of computational search methodologies (evolutionary computation in particular). In the long term, it has significant impact to a broader spectrum of computational intelligence, meta-heuristics, operational research (OR) and even traditional optimisation. At the broadest level, this research explores novel synergies among nature-inspired computation, machine learning, human-computer interaction and high-performance parallel computing, which create strong ties among several EPSRC portfolios (Artificial Intelligence Technology, OR, human-computer interaction) and will address the Cross Information and Communications Technology (ICT) Programme Priority 'Cross-Disciplinarily and Co-Creation' and UKRI cross-organisational themes and programmes. Furthermore, it also addresses other 3 out of the 6 ICT Programme Priorities (none can address all six). Specifically, as the proposed methods place adaptive automation at the heart of the development process, it is directly relevant to the Priority 'Future Intelligent Technologies'. As the domain knowledge from human expertise can be selectively transferred for accelerating new optimisation tasks, thus realising human-in-the-loop, this research will be highly relevant to the Priorities 'People at the Heat of ICT'. Furthermore, data mining and knowledge discovery upon optimised solutions and innovization will address the Priority 'Data Enabled Decision Making'.

This research will directly impact the optimal water system planning and management, which will provide resilient, safe, reliable, and efficient operations for future UK water companies. Furthermore, the programme of research is directly relevant to the government's Industry Strategy, since cost-effective, adaptive computational search is critical to many applications that underpin the Industry Challenge. For example, it is important in tackling challenges in 'Manufacturing and Future Materials' such as the design of light-weight composite materials and managing sustainability across manufacturing systems. Optimisation of complex systems is important to the 'Transforming Construction', 'Prospering from the Energy Revolution' and Clean Growth, e.g., smart building deployments, renewable energy devices. In 'Leading-edge Healthcare', optimisation is essential to accelerate new drug discovery and the improvement of diagnostic tools.

The pathways to impact can be summarised as:

Ensuring academic impacts: (i) present research results at major international conferences in the evolutionary computation, artificial intelligence fields; (ii) publish research results at prestigious international journals; and (iii) organise workshops and disseminate tutorials associated with major conferences to foster an active state of the investigated programme.

Engaging with industries: (i) organise four industry-focused workshops in Exeter to reach out companies who are interested in advanced optimisation techniques in their regular production/operation process, and engage them as research partners to develop case studies; (ii) actively seek opportunities for commercial exploitation of the research outcomes and engage University's Innovation, Impact and Business whenever appropriate; and (iii) offer free technical consultancy on cutting-edge optimisation techniques for a limited period (typically 1-2 days) to SMEs mainly in the south west region.

Engaging with general public: (i) collaborate with the University's Outreach team to host summer schools; and (ii) deliver demonstrations, presentations about research and impacts at University open days.

People and knowledge: (i) PDRAs will gain practical leadership and project management skills by co-supervising PhD students; (ii) research results will be used as study materials in taught undergraduate/postgraduate level modules.