CSIRO Information and Communications Technology Centre, Australia.
Artif Life. 2011 Fall;17(4):293-314. doi: 10.1162/artl_a_00040. Epub 2011 Jul 15.
Small-world networks have been one of the most influential concepts in complex systems science, partly due to their prevalence in naturally occurring networks. It is often suggested that this prevalence is due to an inherent capability to store and transfer information efficiently. We perform an ensemble investigation of the computational capabilities of small-world networks as compared to ordered and random topologies. To generate dynamic behavior for this experiment, we imbue the nodes in these networks with random Boolean functions. We find that the ordered phase of the dynamics (low activity in dynamics) and topologies with low randomness are dominated by information storage, while the chaotic phase (high activity in dynamics) and topologies with high randomness are dominated by information transfer. Information storage and information transfer are somewhat balanced (crossed over) near the small-world regime, providing quantitative evidence that small-world networks do indeed have a propensity to combine comparably large information storage and transfer capacity.
小世界网络是复杂系统科学中最具影响力的概念之一，部分原因是它们在自然发生的网络中普遍存在。人们常常认为这种普遍性是由于其具有高效存储和传输信息的固有能力。我们对小世界网络相对于有序和随机拓扑的计算能力进行了综合研究。为了为这个实验生成动态行为，我们为这些网络中的节点赋予随机布尔函数。我们发现，动力学的有序相（动力学中的低活性）和低随机性拓扑主要由信息存储主导，而混沌相（动力学中的高活性）和高随机性拓扑主要由信息传输主导。信息存储和信息传输在小世界区域附近（交叉）达到某种平衡，为小世界网络确实具有组合相当大的信息存储和传输能力的倾向提供了定量证据。
