Combinatorial Models for the Dynamics of Complex Decision Systems

Combinatorial Models for the Dynamics of Complex Decision Systems

Michael Coombs, Reinhard Laudenbacher, and Karen Schlauch
Physical Science Laboratory, Mathematical Sciences, New Mexico State University


One important factor which determines the dynamics of a decision system is the distribution of ``influence'' in the network, understood quite broadly. This talk will describe an approach to capturing this influence structure with new combinatorial and algebraic tools. The space of possible states of the system does usually not carry any structure beyond that of a directed graph. We enrich this structure by constructing a discrete analog of a vector bundle on this space, that captures the local flow of influence. Algebraic invariants of this discrete vector bundle give structural properties of the network, which are related to its dynamics. We will illustrate these ideas with experimental results for Boolean networks.