Sunday, May 27, 2007

Adventures in Agent Based Modelling

I'm attempting to apply ABM in my PhD to simulate the emergence of new manufacturing capabilities in smaller firms. I purchased a copy of AnyLogic and will be updating the group on my progress. Does anyone know of similar work in this area: using ABM to model organizational capabilities?


Friday, April 13, 2007

Swarm Intelligence

One of my earliest introductions to ideas related to complexity theory.

Monday, April 9, 2007

Many Eyes

This site allows you to create visualizations of your data (social networks, etc) and publish them online.

Out of Control

Free book online on complex systems. Description and link is below:

Out of Control is a summary of what we know about self-sustaining systems, both living ones such as a tropical wetland, or an artificial one, such as a computer simulation of our planet. The last chapter of the book, "The Nine Laws of God," is a distillation of the nine common principles that all life-like systems share. The major themes of the book are:

  • As we make our machines and institutions more complex, we have to make them more biological in order to manage them.

  • The most potent force in technology will be artificial evolution. We are already evolving software and drugs instead of engineering them.

  • Organic life is the ultimate technology, and all technology will improve towards biology.

  • The main thing computers are good for is creating little worlds so that we can try out the Great Questions. Online communities let us ask the question "what is a democracy; what do you need for it?" by trying to wire a democracy up, and re-wire it if it doesn't work. Virtual reality lets us ask "what is reality?" by trying to synthesize it. And computers give us room to ask "what is life?" by providing a universe in which to create computer viruses and artificial creatures of increasing complexity. Philosophers sitting in academies used to ask the Great Questions; now they are asked by experimentalists creating worlds.

  • As we shape technology, it shapes us. We are connecting everything to everything, and so our entire culture is migrating to a "network culture" and a new network economics.

  • In order to harvest the power of organic machines, we have to instill in them guidelines and self-governance, and relinquish some of our total control.

Wednesday, April 4, 2007

Martin Woolf's Review of Complexity in Economics

Standard economics and the 'evolution' thesis can coexist By Martin Wolf

Published: January 16 2007 17:36 Last updated: January 16 2007 17:36

Is the discipline of economics built on sand? Most economists would answer with a resounding "no". But most must also know that the economy is not characterised by perfect foresight and equilibrium, but by trial and error and evolution. That was the intuition of the Austrian economists, Joseph Schumpeter and Friedrich Hayek. But this vision has had next to no influence in the discipline itself.

This gap between how economists think and what economies are is evident to any careful observer. But hitherto nobody has closed the gap between rigorous theory and broad vision. This, argues McKinsey's Eric Beinhocker in a brilliant, thought-provoking and wide-ranging book, published last year, is about to change.* Welcome, he argues, to the world of "complexity economics", computer-based simulations and more realistic assumptions.

Mr Beinhocker has a measure of the complexity of the modern economy - the number of distinct products, or "stock keeping units". In a stone-age culture the number was a few hundred. In today's New York, he suggests, the number may be 10bn. Moreover, not just most of those products but the complex system that invented, designed, produced and sold them is largely the result of just the last 250 years out of 2.5m years of human evolution.

"The economy is a marvel of complexity," he states. "Yet no one designed it and no one runs it." How can such a system have been created? Why has complexity increased over time? Why has so much of the rise in wealth and complexity been so sudden? The answer to these questions can be found, suggests Mr Beinhocker, in understanding that the economy is "a complex adaptive system", which works under the same logic as biological evolution - differentiate, select and amplify.

Conventional economics cannot explain such an evolutionary process, because the science that has provided most of its ideas is not biology, but physics. Conventional economics assumes that human beings are rational, consistent, far-sighted and selfish. But human beings are not desiccated calculating machines. They decide quickly and make predictable mistakes. Their evolutionary history equipped them with the ability to survive in a complex, fast-changing and often dangerous environment.

For any living creature, the evolutionary game involves obtaining resources to live long enough to procreate and rear its young. The economy is humanity's successful effort at obtaining those resources. It has permitted the astounding increase in human numbers and wealth of the past 10,000 years and, above all, of the past 250 years (see charts).

As Mr Beinhocker puts it: "The economy is ultimately a genetic replication strategy. It is yet another evolutionary Good Trick . . . built on the complex Good Trick of big brains, nimble tool-making hands, co-operative instincts, language and culture." So successful is the economy that much of humanity no longer has to focus on staying alive. But our aims are still those of our ancestors: sustenance, shelter, clothing, transport and entertainment with lives built around job, mate, home and children.
How has today's economy evolved? The answer is: through the interaction of three processes. The first is the evolution of physical technology, which spurted after the scientific revolution of the 17th century. The second is the evolution of social technologies, such as money, markets, the rule of law, the corporation and democracy. The third is the evolution of businesses, the entities that live, die and replicate in the economy.
Economic evolution and biological evolution are different: the fact that human beings can plan and adapt makes economic evolution faster and more purposive than biological evolution. But it is still evolution. Also different is what determines fitness. In biology it is survival. In economics the decision-maker was usually a "Big Man" - a chieftain, king or planner. The social technology that changed this was the market economy. It has made the mass of consumers sovereign.
As Mr Beinhocker puts it: "Markets win over command and control, not because of their efficiency at resource allocation in equilibrium, but because of their effectiveness at innovation in disequilibrium." Markets are a hugely powerful evolutionary mechanism: they are innovation machines.
Is this thesis true? Is it useful? Does it replace standard economic analysis? The answers to these questions, I believe, are: yes; yes; and no.
First, today's economy has indeed evolved. Economic evolution is not the same as biological evolution. But it is a member of the general class of evolutionary systems.
Second, this way of thinking is extremely useful. The book explains, for example, why most companies fail to sustain competitiveness. General Electric is the only Forbes 100 company to have both survived and outperformed the market since 1917. The fundamental reason for this is that existing businesses find it far harder to jump into new business niches than new ones. As Mr Beinhocker notes: "Companies don't innovate; markets do." So businesses should think in terms of evolutionary adaptability. But they find this difficult, because they are far better at executing plans than adapting to unforeseen circumstances. That is the price they pay for hierarchy.
Finally, what does "complexity economics" mean for economics? Much less than Mr Beinhocker imagines, I believe. Even such great evolutionary theorists as William Hamilton and Maynard Smith also used equilibrium models. Similarly, the economist's simplification of human motivation is often the only way to make a complex problem tractable. The implications of the ad hoc, computer-based simulations Mr Beinhocker recommends are often difficult to understand. Above all, traditional economics often works: look at the success of inflation targeting or at the benefits of trade.
Just like biologists, economists will have to use different tools for analysing the economic system as a whole from those used for narrower analytical questions, such as the impact of a congestion charge or carbon tax. Even so, the evolutionary perspective is important: competition policy is one compelling example.
The ideas reported in this book will probably change economics much less than the author hopes. But anybody interested in understanding why we are where we are should read it. For me, it was more than the business book of 2006; it was the book of 2006.

*The Origin of Wealth: Evolution, Complexity and the Radical Remaking of Economics (Random House, 2006)