Comparing behavior patterns of swarms that learn using tolerance perceptual near sets

Abstract

The problem considered in this paper is how to measure the degree of nearness of behaviours of swarms that learn. The solution to this problem is set forth a methodology for discovering perceptual granules (i.e., sets of perceptual objects) that are, in some sense, close to each other. A perceptual object is something presented to the senses or knowable by the mind. The basic approach to comparing perceptual objects is inspired by the early 1980s work by Zdzis law Pawlak on the classification of objects. Objects are classified by comparing descriptions of objects stored in information tables. In this work, descriptions of swarm behavior are stored in tables called rough ethograms. A swarm behavior description is defined by means of probe functions (sensor readings) that represent behaviour features. The proposed approach to comparing bahaviours and extracting pattern information in different ethograms takes advantage of recent studies of the nearness of objects and near sets. Behavior patterns are near each other if they have similar descriptions. The contribution of this paper is a framework for determining the nearness of behaviours represented in two or more ethograms.