Distance graphs: from random geometric graphs to Bernoulli graphs and between

We introduce and study random distance graph. A random distance, D(n,g) results from placing n points uniformly at random on the unit area disk and connecting every two points independently with probability g(d), where d is the distance between the nodes and g is the connection function. We give a connection function g(r,a,d) with parameters r and a and show the following: (a) D(n,g(r,a)) captures as special cases both the standard random geometric graph G(n,r) and the Bernoulli random graph B(n,p) (a.k.a. Erdos-Renyi graph). (b) Using results from continuum percolation we are able to bound the connectivity threshold of D(n,g(r,a)), with G(n,r) and B(n,p) as special (previously known) cases. (c) Contrary to G(n,r) and B(n,p), for a wide range of r and α a is, in fact, a "Small World" graph with high clustering coefficient of about 0.5865a and diameter of ({log n}/{log log n}). As opposed to previous Small World models that rely on deterministic sub-structures to grantee connectivity, random distance graphs offer a completely randomized model with a proved bounds for connectivity threshold, clustering coefficient and diameter.