Initial Formulation of an Optimization Method Based on Stigmergic Construction

Initial Formulation of an Optimization Method Based on Stigmergic Construction

Aditya C. Velivelli (Ames Laboratory, USA) and Kenneth M. Bryden (Ames Laboratory, USA)
Copyright: © 2012 |Pages: 18
DOI: 10.4018/978-1-4666-1565-6.ch004


Sign-based stigmergic methods such as the ant colony optimization algorithm have been used to solve network optimization, scheduling problems, and other optimization problems that can be visualized as directed graphs. However, there has been little research focused on the use of optimization methods based on sematectonic stigmergy, such as coordination through collective construction. This paper develops a novel approach where the process of agent-directed stigmergic construction is introduced as a general optimization tool. The development of this new approach involves adopting previous work on stigmergic construction to a virtual space and applying statistical mechanics–based techniques to data produced during the stigmergic construction process. From this a unique procedure for solving optimization problems using a computational procedure that simulates sematectonic stigmergic processes such as stigmergic construction is proposed.
Chapter Preview

Modeling Stigmeric Construction

Assembled structures can emerge from numerous local interactions between the components or agents in a given environment. The motivation behind modeling stigmergy based collective construction is based on observations of paper wasps that build their nests by engaging in reactive, stimulus-response type behaviors. Although they may not have any explicit knowledge about the overall structure of the nest they are building, they can observe the local arrangements of existing nest elements. This observation leads to the insect behavior or response regarding depositing a new nest element. The local patterns that lead the insect to engage in a reactive behavior were termed stigmergic configurations by Theraulaz and Bonabeau (1995), who also performed a computational study of the nest construction of paper wasps. In their study the modeled wasps moved randomly in a three-dimensional lattice and reacted to stigmergic configurations when they were close enough to detect one. Theraulaz and Bonabeau designed the stigmergic configurations or rule sets manually. In a later study by Bonabeau, Guerin, Snyers, Kuntz, and Theraulaz (2000) the rule sets were evolved using a genetic algorithm.

Complete Chapter List

Search this Book: