Novel Bio-Inspired Technique of Artificial Social Cockroaches (ASC)

Abstract

This paper deals about a new bio-inspired algorithm that can be classified in the family of swarm optimization algorithms. The authors' algorithm, which is called Artificial Social Cockroaches (ASC), is inspired from the social behaviour of cockroaches. This inspiration is based on the general phenomenon of real cockroaches that resides in grouping them under the same shelter (place with less lightness) and the way of choosing which shelter and how to get into it. This algorithm has as input a population of artificial cockroaches that will cooperate among them from iteration to another to solve a specific problem using simple rules as the attraction method and the aggregation operators (interaction, individual preference and evaluation). In order to evaluate our algorithm, the authors confronted several experiments facing clustering problem by applying this model on Reuters benchmark and basing on three essential measures: number of Bell, number of Stirling and time complexity.