Complexity Science Leads To Chaos Inc.
from the agent-based-modeling dept
Red Herring has an article looking at the small, but growing, industry surrounding agent-based modeling and complexity science to deal with complicated business issues. The article includes a few examples, including one where Southwest Airlines had a company create an agent-based simulation of their freight shipping procedures, which came up with some easily implementable suggestions that’s saving them millions. I’ve always been fascinated by this sort of thing, but I do wonder how difficult it is to really model these complex systems correctly. While I can see how computers can be useful in playing with the variables to see what comes out – all of those model variables have to be created at first, and if you miss one or two important variables, then the whole model is shot.
Comments on “Complexity Science Leads To Chaos Inc.”
We’ve had Inventory Control Systems and Expert Systems for at least 15 years — this appears to be their latest reincarnation. This is an issue of hardware catching up to the software — with better hardware, we can make more intuitively straightforward software along the lines of SimCity, as opposed to the mathematical charts of old.
I remember IBM tried to upgrade the FAA’s traffic control system in the mid-90s, and it turned into the most spectacular failure in software history — $1 billion lost.
What if software like this proves that the Peace Corps do more harm than good, by increasing the population of backward nations and therefore the risk of war, natural disasters? This kind of software will be a lobbyist’s dream, feeding it only the data they want.