Using the theory and quantitative methods of guided self-organisation, large-scale complex networks and distributed computation, we aim to improve prevention and management of techno-socio-economic and environmental crises.
Modern smart cities, infrastructure and ecosystems are susceptible to abrupt, large-scale and disruptive dynamics. Cascading power failures, traffic disruptions, epidemic outbreaks, financial and housing market crashes, and ecosystem collapses are all manifestations of critical phenomena.
Our aim is to produce new cross-disciplinary methods for understanding and managing complex systems, based on information theory, distributed and high-performance computation, network theory, agent-based simulation, dynamical systems, game theory, machine learning, computational epidemiology, econophysics and systems biology.
This cross-disciplinary research leverages our expertise in engineering and computational sciences and involves collaborations across physics, mathematics, biology and social sciences. Research outcomes have an impact on diverse areas such as:
The Complex Systems group brings together several areas that investigate and innovate with complexity, including the Centre for Complex Systems, the Centre for Translational Data Science and the Centre for Distributed and High Performance Computing.
Our key research areas are guided self-organisation (GSO) and complex networks. We are also involved in several international collaborations related to complex systems.
Self-organisation is the evolution of a system into an organised form, in the absence of external influences. It brings many attractive properties to systems, such as robustness, adaptability and scalability. Self-organising systems can be found almost everywhere: neuronal ensembles self-organise into complex spike patterns; schools of fish change shape in response to approaching predators; and ecosystems develop spatial structures in response to diminishing resources.
The goal of guided self-organisation (GSO) is to leverage the strengths of self-organisation (that is, its simplicity, parallelisation, adaptability, robustness and scalability) while still being able to direct the outcome of the self-organising process. We approach GSO by examining the information processing within complex systems, especially near-phase transitions and critical regimes. We use information-theoretic measures of complexity, criticality, and computation (“information dynamics”) and develop cross-disciplinary methods, relating information theory and thermodynamics (“information thermodynamics”), as well as game theory and complex networks.
Our partners: Associate Professor Manoj Gambhir (IBM Research, Australia), Mr Timothy Germann (Los Alamos National Lab, USA)
The project aims to considerably improve the accuracy and scope of modern computational epidemiological models. It integrates large-scale census datasets and explicitly simulates the entire population down to the level of single individuals, coupled with complex network-based and information flow analysis. The intended outcomes include a more precise and efficient forecasting of critical epidemic dynamics.
Related story: We will find a way to predict the spread of disease
Our partners: Professor Doyne J Farmer (University of Oxford, UK), Mr Paul Ormerod (Volterra Partners, UK), Dr Markus Brede (University of Southampton, UK)
This project aims to improve our understanding of the housing market in Australia by using better modelling, simulation and prediction of the systemic risks and potential crises it faces. We combine datasets from the Australian Bureau of Statistics, census data and mortgage market data into a simulation of individual family homes and their financial tensions, in order to stress test various policies at different scales.
This project integrates research strengths in epidemiological and computational modelling of distributed systems and complex networks, including prediction of critical dynamics during epidemics. It aims to develop a novel computational framework for analysis and modelling of food-borne disease dynamics.
Networks are ubiquitous in today’s world. Communication networks are changing the way we live and interact. Social networks are redefining the ways we keep in touch. Transport networks give us access to the remotest parts of the world. The energy needed for our domestic and industrial use is supplied by electric power networks. Human survival depends on the functioning of a number of biological and ecological networks.
In complex networks the ability to function effectively arises not from individual nodes of that network, but from the way they interact. This means a complex network cannot be completely understood by examining each of its parts in isolation, and that the structure and function of such networks are tightly coupled: function is constrained by structure and structure evolves due to function. Research into the structure, function, evolution, and design of complex networks has wide-ranging applications, from epidemiological modelling to optimising distributed computation.
Our experts: Dr Joseph Lizier
Our partners: Professor Michael Wibral (Goethe University, Frankfurt, Germany), Ms Viola Priesemann (Max Planck Institute for Dynamics and Self-Organisation, Göttingen, Germany).
This project seeks to develop a general understanding of how network connectivity and dynamics are related by:
Our experts: Professor Mikhail Prokopenko, Dr Eduardo Altmann, Professor Deborah Bunker, Professor Roland Fletcher, Dr Joseph Lizier, Professor Richard Miles, Dr Ramil Nigmatullin, Dr Daniel Penny, Dr Somwrita Sarkar, Mr Tony Sleigh
Our partner: Paul Ormerod (Volterra Partners, UK)
The CRISIS research program is developing a novel cross-disciplinary framework for analysis, modelling and design of adaptive urban systems resilient to stresses, using advanced techniques from complex systems, network science, agent-based computational modelling, and dynamical systems. To evaluate this general framework we selected two settlement areas with distinct features and different histories yet of similar physical size and urban landscape: the Greater Sydney area in the 20th and 21st centuries and Greater Angkor area in the 13th and 14th centuries.
Security concerns are increasingly at the forefront in today’s world. A complex network’s topology plays an important role in determining how the network can resist random and targeted attacks. It has been shown that scale-free networks display high resilience to random attacks, and yet these networks are vulnerable to targeted attacks. Is it possible to design network topologies so that they display high levels of tolerance to both types of attacks? This project attempts to define robust measures for complex networks that are particularly useful in sustained attack scenarios, and test their effectiveness in various domains, including computer networks, the Internet and the World Wide Web, social networks, and biological networks inside organisms.
Our key collaborators in Australia include DST Group, CSIRO, IBM Research, and various state and federal government agencies. We also actively engage with the Santa Fe Institute for Complex Systems, Los Alamos National Laboratory and Harvard University in the United States, the University of British Columbia in Canada, the Institute for New Economic Thinking at Oxford University, other universities in the United Kingdom and several Max Planck institutes in Germany.
Our partner: Dr Dominique Chu (University of Kent, UK)
The overarching aim of this project is to develop a theory of biological and cellular computing. This will lead to an information-theoretic assessment of the distributed computation in physical systems and an understanding of biological computation under the information thermodynamics framework. The project is related to another grant from The Royal Society for the Theo Murphy Discussion Meeting: ‘Toward a Computational Theory of Life’ (2018).
Our partners: Mr Guy Theraulaz (Université Paul Sabatier, Toulouse, France), Ms Rosalind X. Wang (CSIRO)
We develop and apply a rigorous information-theoretic framework for detecting and measuring predictive information flows during collective motion within a school of fish. We also explore the role of spatial dynamics in generating the influential interactions that carry the information flows.