لطفا منتظر باشید ...
Messengers and Buildings
To separate our discussion of complexity from an analysis of games, let's look at an example of a system that is not a game. Our sample system is a communications network, a fictional city where a messenger has to deliver messages between buildings. In the most straightforward version of this system, there are only two buildings. Each building generates messages that are addressed to the other building; the messenger has to run between them.
On the whole, this is not a complex system. Although the movement of the messenger and the changing states of the buildings makes it is more complicated than our example of a table, the behavior of the system is set in a fixed, predictable pattern. The messenger will oscillate back and forth between the two buildings. Even if the two buildings generate messages at random intervals, so that the messenger never knows when and where the next message will appear, this does not add substantial new states for the system as a whole. What happens if we add new elements and relationships that push the system toward a higher degree of complexity? Imagine that there are now 10 messengers trying to deliver messages between 50 buildings, randomly positioned on a grid. The different distances between the buildings determine how long the messengers travel. This is important, because we are trying to maximize efficiency: if a message goes undelivered for too long, the communication system has failed to function properly. With the addition of these elements, the system suddenly increases substantially in complexity.
To appreciate the added complexity, consider writing a computer program to manage the behavior of the messengers. In the two-building system, the program would be very simple, merely shuttling the single messenger between the two buildings. But with 10 messengers and 50 buildings, it becomes a more complex task. If a messenger delivers her last message at a building and there isn't a new message to pick up, where does she go next? This decision should be based partially on which buildings are closest to the messenger, but it should also take into account which buildings other messengers have recently visited; buildings that haven't been visited recently are most likely to have an undelivered message.
Or consider what should happen when a messenger picks up several pieces of mail that need to go to buildings in different parts of the city. What is the logic that determines how to deliver the messages? What if the messengers have the ability to "meet" in-between buildings so that they can swap messages to work more efficiently? How, when, and where will the messengers rendezvous? Consolidating deliveries to the same building by swapping messages may make the system more efficient, but if messengers meet too often, the overall efficiency of the system might go down. The functioning of the communication system is no longer a simple, predictable oscillation between two states. It has achieved a fundamentally new degree of complex, compounded interrelationships. And if that scenario isn't complex enough, we could add additional variables. For instance, what if some buildings have special relationships with other buildings (a lot of mail passes between the two of them) and that these relationships change over time? Or perhaps the messengers aren't delivering mail between "buildings," but instead they are delivering to "companies." The companies keep changing their building address, so that their physical locations in the grid of the city constantly shift. Or perhaps messengers never have a bird's-eye view of the entire city, but only gain information about the layout and behavior of the city from their own experience and by sharing information with other messengers. Writing a computer program to simulate all of this would be more challenging than our first scenario. In fact, the sample system highlights some of the problems of information transmission and social behavior modeled in complexity research. What makes our communication system complex? The many dynamic and contingent interrelationships among the parts push the system across the "complexity barrier" Campbell describes. The initial version of the system, with two buildings and one messenger, was primitive and mechanistic by comparison. But the more complex versions of our city resemble an ecosystem filled with intelligent agents. We articulate this conceptual distinction between a simple and a complex system further in the examples that follow.
