• Contribute
• What's New
• Submit News

Introduction

It is an interesting exercise to think of a city as a system. At a high level, cities are made up of various components that take inputs and produce outputs. One example might be turning raw materials and human labour into buildings and infrastructure. But the analogy can indeed be taken much further. Cities produce by-products such as waste and pollution. They consist of many subsystems, such as the transport network and economic systems, and interestingly cities can go wrong or produce unexpected results. They interact with other cities, just as some systems interact with others, and in many instances cities will adapt to changing conditions in their neighbouring cities. This all adds up to a huge level of complexity.

A city is also a very dynamic entity. It is fair to state that the quantity and nature of the inputs and outputs change over time, and that cities are operating in a changing physical, economic and social environment. Much more fundamentally, the people who make up this system posses both intelligence and their own high level goals to pursue. This is in marked contrast to an engineered physical system, or indeed a biological social system such as an ant colony, in which the component parts have very low level goals, and know very little about the higher level environment that they are operating in. In all of these instances though, the system components comes together to produce useful work regardless of the autonomy of the individual agents.

Complex Adaptive Systems are best thought of as a subset of all systems that display specific characteristics. In the remainder of this essay I discuss some of these characteristics, and indicate how cities may demonstrate these same features. I believe that this is a very useful exercise, as establishing this link allows us to draw upon a large body of complex adaptive systems research, experience and expertise that could be used in city planning, development, human organisation and urban evolution.

Nature of agents

Complex Adaptive Systems are often made up from a large number of heterogeneous agents that interact with one another in non linear ways. As this large number of varied agents interact, the overall behaviour of the system becomes very difficult to predict or model. The word Complex is actually referring to this interaction complexity, as opposed to any specific intelligence or attributed to the agents.

With regards to cities, it is clear that humans interact with one another in a multitude of ways. They form relationships and families, exchange goods and services, form groups, steal or hurt one another, and work and socialise together. Each human is different, and each wants different things. As humans move through cities in a typical day these interactions are in line with the complex adaptive system definition, in that we have no hope of predicting the effect and nature of these interactions. In addition, we also see high level patterns emerge from these same interactions, such as crowding, violence, traffic jams or meetings of people with common interests. None of the individual agents even considers the role that they are playing in these phenomena, and none of them could represent a traffic jam alone. These are all important issues that we would like to gain insight into.

To compound this issue, humans are not the only agents acting within a city. Businesses and government organisations can be seen as a cohesive whole that possess their own goals and interaction models, and sometimes these organisations are in conflict or competition, just like the human agents. The human agents and the organisational agents will then of course interact with one another, by for instance businesses and humans influencing the relevant size of each population through employment patterns.

Subsystems

Complex Adaptive Systems may be made of many subsystems or categories of agent that interact with and affect one another. As well as the large number of distinct agents typically found in these systems, the multiple subsystems add infinitely more layers of complexity as they influence one another.

Cities generally posses a transport network, a physical infrastructure, and a supporting business environment for organisations to interact. Economies, communities and societies can also be considered as subsystems operating within the larger city system. It’s clear that each of these have an important influence on one another and that in some instance drive one another or meet in conflict. For instance, trends in society subsystems result in the emerge of business

Evolution and co-evolution

These are two properties widely found in Complex Adaptive Systems. Where agents can adapt as individuals, we will often see trends of adaptation throughout the system as a whole. In some instances, populations evolve against one another,

Cities will essentially evolve from state to state as a result of changing behaviour patterns of agents. Cities as a whole evolve by changing in size and prosperity, and clear cycles in wealth and condition of cities can be observed over many years.

All of these changes take place on a smaller scale within the context of a single city. Communities grow in strength and evolve into different districts. Cycles will occur where districts become unfashionable and subsequently nice places. Groups such as police and criminals, students and teachers, and groups of business will co-evolve, driving each others direction of development

Dynamic and Adaptive

Complex Adaptive Systems are typically dynamic, in that they face changing environments and inputs. These systems naturally adapt to changes that they are confronted with, and will often fit to the idiosyncrasies of their operating environment.

Cities operate under changing conditions. Populations, number of businesses, weather, pollution, the physical environment, immigration and traffic all fluctuate drastically over very short spaces of times. Cities however are robust. As a rule, it is not the case that when some people leave, a dramatic cycle of population and subsequent business collapse ensues. Instead, more business will move in to naturally meet demand. Likewise, transport and network usage will often converge on an appropriate level of use. It’s rare that humans have to interfere with this process.

Emergence

Typically, Complex Adaptive Systems display emergence. Emergence is referring to high level patterns emerging from low level rules and events that might not have been anticipated. As Complex Adaptive Systems are usually constituted by heterogeneous agents, we can often describe the system in terms of their useful emergent behaviour, as opposed to what the individuals are doing. An example might be some of the algorithms that we have observed ants and termites to follow. As a rule, these algorithms are much more useful than a detailed understanding of the local level chemical based interactions between the insects.

Emergence can be observed at many levels of a city. Traffic patterns are typical examples, whereby all individual agents are following simple local rules whilst trying to minimise their own journey time. An example is how traffic jams occur because of clearly identifiable rules followed by internal drivers, such as not keeping a consistent distance from preceding cars. Another interesting example of emergence is that of communities, which form based on simply being in the same location to one another.

Chaotic properties

Complex Adaptive Systems are often very hard to predict. They are often very sensitive to their starting conditions, and it is sometimes the case that a small change in operating conditions or agent strategies can be compounded into an entirely different behaviour at the system level.

Cities often display chaotic properties. The initial state, such as the physical location or constituents of the population may have a massive bearing in how the system goes on to evolve. Likewise, small changes, such as the addition of an unsavoury business or human may go on to produce relatively wide ranging effects. As stated, the system will often keep moving, but relative to their size

Conclusion

True Complex Adaptive Systems display a number of very attractive properties. They are typically very robust, as it’s not the individual components that are producing the useful work, but the interactions between agents. The adaptive properties would make them very attractive solutions where they can be used effectively to solve some real world problems. Emergence is also always attractive, as it presents the idea of getting something for nothing. The idea of systems evolving into superior solutions is also obviously desirable.

Fortunately, cities appear to fit into the definition of complex adaptive systems in many ways. As stated, successfully establishing this allows us to draw upon a wide range of research and knowledge relating to complex adaptive systems, and perhaps allow us to consider cities in a more abstract and quantitative way.