Small-world network

Small-world networks tend to contain cliques, and near-cliques, meaning sub-networks which have connections between almost any two nodes within them. This follows from the defining property of a high clustering coefficient. Secondly, most pairs of nodes will be connected by at least one short path. This follows from the defining property that the mean-shortest path length be small. Several other properties are often associated with small-world networks. Typically there is an over-abundance of hubs – nodes in the network with a high number of connections (known as high degree nodes). These hubs serve as the common connections mediating the short path lengths between other edges. By analogy, the small-world network of airline flights has a small mean-path length (i.e. between any two cities you are likely to have to take three or fewer flights) because many flights are routed through hub cities. This property is often analyzed by considering the fraction of nodes in the network that have a particular number of connections going into them (the degree distribution of the network). Networks with a greater than expected number of hubs will have a greater fraction of nodes with high degree, and consequently the degree distribution will be enriched at high degree values. This is known colloquially as a fat-tailed distribution. Graphs of very different topology qualify as small-world networks as long as they satisfy the two definitional requirements above.

Network small-worldness has been quantified by a small-coefficient, ${\displaystyle \sigma }$, calculated by comparing clustering and path length of a given network to an equivalent random network with same degree on average.[6][7]

${\displaystyle \sigma ={\frac {\frac {C}{C_{r}}}{\frac {L}{L_{r}}}}}$

if ${\displaystyle \sigma >1}$ (${\textstyle C\gg C_{r}}$ and ${\textstyle L\approx {L_{r}}}$), network is small-world. However, this metric is known to perform poorly because it is heavily influenced by the network's size.[8][9]

Another method for quantifying network small-worldness utilizes the original definition of the small-world network comparing the clustering of a given network to an equivalent lattice network and its path length to an equivalent random network. The small-world measure (${\displaystyle \omega }$) is defined as[8]

${\displaystyle \omega ={\frac {L_{r}}{L}}-{\frac {C}{C_{\ell }}}}$

Where the characteristic path length L and clustering coefficient C are calculated from the network you are testing, C_ℓ is the clustering coefficient for an equivalent lattice network and L_r is the characteristic path length for an equivalent random network.

Still another method for quantifying small-worldness normalizes both the network's clustering and path length relative to these characteristics in equivalent lattice and random networks. The Small World Index (SWI) is defined as[9]

${\displaystyle {\text{SWI}}={\frac {L-L_{\ell }}{L_{r}-L_{\ell }}}\times {\frac {C-C_{r}}{C_{\ell }-C_{r}}}}$

Both ω′ and SWI range between 0 and 1, and have been shown to capture aspects of small-worldness. However, they adopt slightly different conceptions of ideal small-worldness. For a given set of constraints (e.g. size, density, degree distribution), there exists a network for which ω′ = 1, and thus ω aims to capture the extent to which a network with given constraints as small worldly as possible. In contrast, there may not exist a network for which SWI = 1, the thus SWI aims to capture the extent to which a network with given constraints approaches the theoretical small world ideal of a network where C ≈ C_ℓ and L ≈ L_r.[9]

Small-world properties are found in many real-world phenomena, including websites with navigation menus, food webs, electric power grids, metabolite processing networks, networks of brain neurons, voter networks, telephone call graphs, and airport networks.[10] Cultural networks[11] and word co-occurrence networks[12] have also been shown to be small-world networks.

Networks of connected proteins have small world properties such as power-law obeying degree distributions.[13] Similarly transcriptional networks, in which the nodes are genes, and they are linked if one gene has an up or down-regulatory genetic influence on the other, have small world network properties.[14]

In another example, the famous theory of "six degrees of separation" between people tacitly presumes that the domain of discourse is the set of people alive at any one time. The number of degrees of separation between Albert Einstein and Alexander the Great is almost certainly greater than 30[15] and this network does not have small-world properties. A similarly constrained network would be the "went to school with" network: if two people went to the same college ten years apart from one another, it is unlikely that they have acquaintances in common amongst the student body.

Similarly, the number of relay stations through which a message must pass was not always small. In the days when the post was carried by hand or on horseback, the number of times a letter changed hands between its source and destination would have been much greater than it is today. The number of times a message changed hands in the days of the visual telegraph (circa 1800–1850) was determined by the requirement that two stations be connected by line-of-sight.

Tacit assumptions, if not examined, can cause a bias in the literature on graphs in favor of finding small-world networks (an example of the file drawer effect resulting from the publication bias).

It is hypothesized by some researchers, such as Barabási, that the prevalence of small world networks in biological systems may reflect an evolutionary advantage of such an architecture. One possibility is that small-world networks are more robust to perturbations than other network architectures. If this were the case, it would provide an advantage to biological systems that are subject to damage by mutation or viral infection.

In a small world network with a degree distribution following a power-law, deletion of a random node rarely causes a dramatic increase in mean-shortest path length (or a dramatic decrease in the clustering coefficient). This follows from the fact that most shortest paths between nodes flow through hubs, and if a peripheral node is deleted it is unlikely to interfere with passage between other peripheral nodes. As the fraction of peripheral nodes in a small world network is much higher than the fraction of hubs, the probability of deleting an important node is very low. For example, if the small airport in Sun Valley, Idaho was shut down, it would not increase the average number of flights that other passengers traveling in the United States would have to take to arrive at their respective destinations. However, if random deletion of a node hits a hub by chance, the average path length can increase dramatically. This can be observed annually when northern hub airports, such as Chicago's O'Hare airport, are shut down because of snow; many people have to take additional flights.

By contrast, in a random network, in which all nodes have roughly the same number of connections, deleting a random node is likely to increase the mean-shortest path length slightly but significantly for almost any node deleted. In this sense, random networks are vulnerable to random perturbations, whereas small-world networks are robust. However, small-world networks are vulnerable to targeted attack of hubs, whereas random networks cannot be targeted for catastrophic failure.

The main mechanism to construct small-world networks is the Watts–Strogatz mechanism.

Small-world networks can also be introduced with time-delay,[16] which will not only produce fractals but also chaos[17] under the right conditions, or transition to chaos in dynamics networks.[18]

Soon after the publication of Watts–Strogatz mechanism, approaches have been developed by Mashaghi and co-workers to generate network models that exhibit high degree correlations, while preserving the desired degree distribution and small-world properties. These approaches are based on edge-dual transformation and can be used to generate analytically solvable small-world network models for research into these systems.[19]

Degree–diameter graphs are constructed such that the number of neighbors each vertex in the network has is bounded, while the distance from any given vertex in the network to any other vertex (the diameter of the network) is minimized. Constructing such small-world networks is done as part of the effort to find graphs of order close to the Moore bound.

Another way to construct a small world network from scratch is given in Barmpoutis et al.,[20] where a network with very small average distance and very large average clustering is constructed. A fast algorithm of constant complexity is given, along with measurements of the robustness of the resulting graphs. Depending on the application of each network, one can start with one such "ultra small-world" network, and then rewire some edges, or use several small such networks as subgraphs to a larger graph.

Small-world properties can arise naturally in social networks and other real-world systems via the process of dual-phase evolution. This is particularly common where time or spatial constraints limit the addition of connections between vertices The mechanism generally involves periodic shifts between phases, with connections being added during a "global" phase and being reinforced or removed during a "local" phase.

Small-world networks can change from scale-free class to broad-scale class whose connectivity distribution has a sharp cutoff following a power law regime due to constraints limiting the addition of new links.[21] For strong enough constraints, scale-free networks can even become single-scale networks whose connectivity distribution is characterized as fast decaying.[21] It was also shown analytically that scale-free networks are ultra-small, meaning that the distance scales according to ${\displaystyle L\propto \log \log N}$.[22]

• Barabási–Albert model – algorithm for generating random networksPages displaying wikidata descriptions as a fallback
• Climate as complex networks – Conceptual model to generate insight into climate science
• Dual-phase evolution – Process that drives self-organization within complex adaptive systems
• Dunbar's number – Suggested cognitive limit important in sociology and anthropology
• Erdős number – Closeness of someone's association with mathematician Paul Erdős
• Erdős–Rényi (ER) model – Two closely related models for generating random graphs
• Local World Evolving Network Models
• Percolation theory – Mathematical theory on behavior of connected clusters in a random graph
• Network science – Academic field - mathematical theory of networks
• Scale-free network – Network whose degree distribution follows a power law
• Six degrees of Kevin Bacon – Parlour game on degrees of separation
• Small-world experiment – Experiments examining the average path length for social networks
• Social network – Social structure made up of a set of social actors
• Watts–Strogatz model – Method of generating random small-world graphs
• Network on a chip – Electronic communication subsystem on an integrated circuit – systems on chip modeled on small-world networks
• Zachary's karate club

• Dynamic Proximity Networks by Seth J. Chandler, The Wolfram Demonstrations Project.
• Small-World Networks entry on Scholarpedia (by Mason A. Porter)