Keith Decker and Victor Lesser
The design of organizations or other coordination mechanisms for groups of computational agents, either interacting with one another or with people, depends crucially on the task environment of which they are a part. Such dependencies include the structure of the environment (the particular kinds and patterns of interrelationships that occur between tasks) and the uncertainty in the environment (both in the a priori structure of any episode within an environment and in the outcomes of an agent’s actions). Designing organizations also depends on properties of the agents themselves--but this has been studied more thoroughly by other researchers. The central idea is that the design of coordination mechanisms cannot rely on the principled construction of agents alone, but must rely on the structure and other characteristics of the task environment--for example, the presence of uncertainty and concomitant high variance in a structure. Furthermore, this structure can and should be used as the central guide to the design of coordination mechanisms, and thus must be a part of any comprehensive theory of coordination. This working paper will briefly describe our modeling framework, TASMS, for representing abstract task environments. We will also briefly describe a family of domain-independent, team-oriented coordination algorithms called Generalized partial Global Planning (GPGP). Having a family of algorithms allows us tailor the algorithm to the environment (or even a specific situation). We will give an example of an analysis inspired by Burton and Obel’s work on organizational structure and technology decomposability.