|
/vol08/halloy03/ | © Copyright 2000 | |||
| Volume 08 | Received: Accepted: |
15 Jul 2000 01 Nov 2000 |
|||
|
Effects on system structure, diversity and stability of the distance exponent in a resource attraction model
Halloy, S. R. P. |
|
| Abstract | |
| It is well known that diversity, connectivity, stability and productivity of complex systems such as ecosystems and economic systems are related in ways which are still a matter of debate. The magnitude of interaction between agents (or species) is a function of their mass, and is an inverse function of "distance" to a given power (the distance exponent, DE). This paper explores the behaviour of system measurements given a range of DE from -64 to +64 using the resource attraction model (RAM) over 20 periods. This is compared to null models with no distance, and with no distance and no mass effects. The highest distance exponent values (DE>4) resulted in the highest numbers of agents (>68), with the lowest extinction rates (20-22%) and the closest approximation to a lognormal. The mean coefficient of variation (CV) of agents in the last ten periods was 21-22% while the total agent richness CV was 0.9 to 1.4 %. Lower DE values led to lower species richness and stability. For DE=0, the number of agents goes down to a minimum, total mass reaches a maximum (a "monoculture"), and there is high turnover and instability with high distance from the lognormal attractor. The results are context dependent, but, within the range of values explored, the shape of the curves remains, i.e. the singular peak observed close to a zero exponent is a universal feature. This method provides a pathway to further explore the behaviour of the important distance feature in agent-interaction models. This paper defines the theoretical limits to what is "enough" interactions for complex dynamics and in the same process determines the limits to the lognormal patterns which characterise unconstrained complex systems. | |
| Full Text |
|
|