by Matt Garrod
This year the group’s summer meeting was held at Warwick University’s Complexity Centre. Approximately 30 students and researchers congregated in the Complexity Centre to escape the summer heat. As one of the speakers pointed out later in the afternoon, the meeting’s theme brought together two of the least well defined terms in science; the result of this was a relatively diverse crowd and plenty of opportunity to discuss ideas from apparently unrelated fields.
The first talk of the day was given by UCL’s Professor Michael Batty. Michael Batty is somewhat of a pioneer of the scientific movement to understand and quantify the properties and behaviours of cities. He has been applying computer based models to the field since the 1970’s, well before the concept of “complexity” entered the scene. He currently heads the Centre for Advanced Spatial Analysis at UCL, who make use of sophisticated computational and mathematical techniques to better understand how cities function. This is becoming an increasingly important area of research as the world’s population becomes more urbanized.
One of the major problems in the field is of deciding how to define a city geographically. As Batty explained, the task of identifying city limits is straightforward for his colleagues in the US, since many of the major cities are right in the middle of the desert. However, in many European countries such as the UK, the task is more difficult since sprawling suburbs routinely approach smaller satellite towns, making it hard to identify exactly where the “city” ends.
Recent work by Batty and his colleagues has applied ideas from the burgeoning field of network science in order to identify city boundaries with more certainty. Their methods involve representing the UK’s streets as a network. In this representation intersections such as crossroads and roundabouts become “nodes,” while the roads which connect them become “edges”. Using this representation and some ingenious computer algorithms Batty’s team have managed to identify “city-limits” for UK cities. Interestingly, the techniques even managed to identify other features of the United Kingdom, such as the “North-South Divide.” In addition to this, Batty also spoke about how other properties of cities, such as average income, can scale with city size.
The work by Batty and his colleagues is beginning to close in on some of the universal features of cities, showing us that they are not just chaotic urban jungles. One major challenge that does remain however, is of how to extend the techniques used to other countries, where data about road networks can be less accessible or even nonexistent.
The next talk was given by Professor Robert Mckay, Professor of Mathematics at the University of Warwick and director of the Centre for Complexity Science. The talk focused on a recent report on water abstraction in the UK, in which the authors made use of agent-based models in order to support their conclusions. Mckay briefly discussed some of the key mathematical concepts that policy makers should be aware of when making use of agent-based models. For instance, asking questions such as “Does the agent based model predict unique statistical behaviour?” might allow policy makers to develop a greater understanding of exactly what their models are telling them.
Dr Nick McKullen, a lecturer in the department of Architecture and Civil Engineering at the University of Bath, presented the first of several talks on the social aspects of achieving sustainability. He began by discussing how the probability that a household invests in solar panels often depends on whether or not their neighbours possess them.This is a prime example of how the effects of social influence can be more important than the quality of a technology when considering how many people actually invest in it. He spoke in general about how we can use diffusion along social networks in order to model the uptake of green technologies. It appears as if an important part of achieving a sustainable future will rely on understanding the social processes which influence others to copy others behaviour.
In an afternoon which mainly focused on the social aspects of achieving sustainability, Dr Anna Maria Cherubini’s talk on desertification added a bit of variety to the line up. Her talk focused an agent based model of the process which sees the transformation of arid landscapes containing vegetation to lifeless deserts. By understanding the process of desertification it is hoped that we might be able identify early warning signs. These will then allow us to identify regions at risk of change.
The second keynote speaker of the day was Dr Martino Tran, Assistant Professor at the School of Community and Regional Planning at the University of British Columbia and an associate at the Environmental Change Institute at the University of Oxford. His work focuses on the understanding and management of risk and sustainability in cities.
Tran began by pointing out that a large amount of research into sustainability and risk is focused on technological solutions rather than societal or behavioural issues. To fully understand how complicated systems such as the energy industry work we must understand the coevolution between infrastructure, technology and people. In order to understand the behaviour of these systems mathematically researchers generally build mathematical models at three different levels of abstraction. Firstly, “System level” models which study the average behaviour of the system, these include models of diffusion. Secondly models of agent behaviour which include game theoretic and decision theoretic models. Finally models at the level of individual agents, these might involve the study of interactions along networks. By studying social systems at various different levels of abstraction Tran and his colleagues hope to understand some of triggers of mass behavioural change.
Towards the end of the talk Tran gave some advice for physical scientists hoping to make an impact in the environmental sciences. Many physical scientists often strive to understand the world through precise laws and relations; akin to those of Newtonian physics. However, in the environmental sciences, where the systems under scrutiny are universally complicated, it is often necessary for researchers to settle for the more modest goal of identifying the possible scenarios likely to occur. This prescription probably applies to any physical scientists moving towards the study of “complex” systems.
Discussion of society’s interaction with the rest of the world was continued by Professor Bruce Edmond, director of the centre for policy modelling at Manchester Metropolitan University. His recent research has focused on modelling the interaction between social and ecological systems. In recent work he and his colleagues have created a model ecosystem composed of many species. They then investigated what happens to the model ecosystem when a human-like species is introduced. The complexity of the model meant that great variation in the possible outcomes was observed. However, some of the results were not encouraging since the introduction of human-like species generally lead to a loss of biodiversity in the model ecosystems.
The final talk of the day was given by Dr Samuel Johnson, Assistant Professor at the University of Warwick. The lecture focused on one of the key issues that must be confronted in order to tackle global problems such as climate change. The so called “Tragedy of the Commons” occurs when selfish behaviour by individuals in a society comes to the detriment of the group as a whole. He presented some research which was inspired by the staple model of Ferromagnetism in Physics – the Ising Ferromagnet. After presenting his model he made some suggestions about how we may be able to achieve cooperation between heterogeneous nations in the real world. This sparked discussion of exactly what he was proposing and how the organisation is arranged. Such debate is probably inevitable when the topic strays so close to global politics.