Quantitative measures of organization for multiagent systems

A set of "information theoretic" measures has been developed to quantify the degree of constraint inherent in the organization of a multiagent system and the ascendancy of evolving system has a propensity to increase.

Abstract

A set of "information theoretic" measures has been developed to quantify the degree of constraint inherent in the organization of a multiagent system. Separate measures can be provided to quantify spatial organization, trophic organization and, more generally, the overall structure of interactions. The additive character of these quantities allows them to be distributed in various fashions among species and places in a way that allows one to assign an "Importance Index" to those taxa and places. In addition, a measure to gauge the degree of adaptation of a species to a particular environment is proffered. The proposed measures allow one to formulate the following hypotheses in quantitative fashion: (1). that any disturbance of an ecosystem at a location associated with a high spatial Importance Index will exert a greater impact on the population dynamics than will a similar disturbance aimed at a place where the values of these indexes are lower; (2). that any disturbance in an ecosystem affecting a particular species with high individual Importance Indexes will cause a greater impact on the overall population dynamics than will a disturbance aimed at a species with a lower values of these indexes; (3). that the ascendancy of evolving system has a propensity to increase. The precise quantitative formulation of these hypothesis would permit them to be tested via multiagent simulation. Estimating the probablities pertaining to these hypotheses presents a number of problems that merit discussion.