Title :
Stationarity and nonstationarity in a self-organizing information system with primitive intentions
Author :
Kugler, Peter N.
Abstract :
Summary form only given. Nest building by termites is considered as an example to acquire insight into the nature of self-organizing information systems. The behavior of the insects as atomistic units is organized by the time evolution of relatively stationary attractors in the pheromone diffusion field arising from the insects´ waste deposits. A change in the attractor layout induces an instability in the pheromone flow pattern, driving the system to a state of greater order, as instability begets self-organization. As a result, the nest building evolves through stages of random depositing, pillar construction, arch construction and dome construction. Nonstationarity is present in the form of time-dependent variations in the perceptual threshold of the individual insect. When the waste is deposited, the insect loses its pheromone affinity and ceases to be oriented by the diffusion field: the affinity returns with the accumulation of new waste. Consequently, sensitivity to the qualities of the field vary dynamically as a function of simple internal states. This overall insects-environment system of evolving global stationarities and local nonstationarities has been analyzed through the methods of nonlinear dynamics and simulated. Analysis and simulation suggest a number of potentially very general design principles for self-organizing information systems, among them the understanding that the attractors of such systems are emergent cross-scale symmetries
Keywords :
behavioural sciences; cognitive systems; self-adjusting systems; zoology; arch construction; dome construction; emergent cross-scale symmetries; equilibrium; insect behaviour; instability; internal states; nonlinear dynamics; perceptual threshold; pheromone diffusion field; pheromone flow pattern; pillar construction; random depositing; sensitivity; stationarity; stationary attractors; termite nest building; time evolution; time-dependent variations; waste deposits;
Conference_Titel :
Neural Networks, 1990., 1990 IJCNN International Joint Conference on
Conference_Location :
San Diego, CA, USA
DOI :
10.1109/IJCNN.1990.137912