Behavioral aggregation within complex situations: a case study involving dynamic equilibria
Shankar A. Rajamoney, Sang Koo · National Conference on Artificial Intelligence · 1991
The analysis of large complex situations poses difficult problems for qualitative reasoning due to the complexity of reasoning from first principles and the proliferation of ambiguities. Abstraction is a promising Bolution to these problems. In this paper, we study a type of abstraction, behavioral aggregation--the process of grouping a set of individual entities that collectively behave as a unit. In particular, we show how to build aggregate models of situations involving dynamic equilibria and how to reason about their behavior. Finally, we demonstrate, through several examples, the benefits of reasoning at the aggregate level: a reduction in the complexity of reasoning and a compact, easily interpretable, description of the behavior.