Evolving Information Processing Organizations

The organization of information processing resources is a central question in economic, organizational, and computational theory. Recent work by Radner (1992) and others has developed a simple theoretical framework and some useful formal mathematical results about the behavior of such systems. Here, we follow a complementary computational approach that allows us to pursue questions concerning the impact of coordination and various exogenous conditions facing the organization. We find that organizations demonstrate ``order for free,'' that is, given a simple structural framework and a set of standard operating procedures, even randomly generated organizations imply well-defined patterns of behavior. Using a genetic algorithm, we also show that simple evolutionary processes allow organizations to ``learn'' better structures.