Recursive graphs with small-world scale-free properties
Extracted source text ยท arXiv cond-mat/0402033
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%
%% Recursive Graphs with Small-World Scale-Free Properties
%% Comellas, Fertin & Raspaud
%% REVTeX.4.0 document
%%
%% v 0.4 24-NOV-02
%% v 1.0 1/MAR/03 Sent
%% v 1.2 10/MAR/03 FC Introduced new diam
%% v 1.2 19/MAR/03 FC corrections
%% v 1.3 26/MAY/03 FC more corrections
%% v 2.0 1/AUG/03 Sent 1st revision
%% v 2.1 6-8/DEC/03 FC Changing introduction
%% 9/DEC/03 FC Added Table captions. Changed discussion.
%% 13/DEC/03 FC Added references
%% v 3.0 12/JAN/04 FC changed figures and captions as requested
%% v 4.0 1/APR/04 FC corrections by PRE
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
% as well as the rest of the prerequisites for REVTeX 4.0
%
% See the REVTeX 4 README file
% It also requires running BibTeX. The commands are as follows:
%
% 1) latex cl-trees.tex
% 2) bibtex cl-trees
% 3) latex cl-trees.tex
% 4) latex cl-trees.tex
%
\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
%\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
% Some other (several out of many) possibilities
%\documentclass[preprint,aps]{revtex4}
%\documentclass[preprint,aps,draft]{revtex4}
%\documentclass[prb]{revtex4}% Physical Review B
\usepackage{graphicx}% Include figure files
\usepackage{dcolumn}% Align table columns on decimal point
\usepackage{bm}% bold math
%\nofiles
\begin{document}
\title{Recursive graphs with small-world scale-free properties}% Force line breaks with \\
\author{Francesc Comellas}
\affiliation{
Departament de Matem\`atica Aplicada IV, EPSC, Universitat
Polit\`ecnica de Catalunya\\
Avinguda del Canal Ol\'{\i}mpic s/n, 08860
Castelldefels, Barcelona, Catalonia, Spain
}%
\homepage{http://www-mat.upc.es/~comellas}
\email{comellas@mat.upc.es}
\author{Guillaume Fertin}
\affiliation{
IRIN, Universit\'e de Nantes, 2 rue de la Houssini\`ere,
BP 92208, 44322 Nantes Cedex 3, France
}%
\homepage{http://www.sciences.univ-nantes.fr/info/perso/permanents/fertin/}
\email{fertin@irin.univ-nantes.fr}
\author{Andr\'e Raspaud}
\affiliation{
LaBRI, Universit\'e Bordeaux I, 351 Cours de la Lib\'eration,
33405 Talence Cedex, France
}%
\homepage{http://www.labri.fr/Perso/~raspaud/}
\email{raspaud@labri.fr}
\date{}
\date{Rec. 26 November 2002; revised manuscript rec. 15 December 2003; publ. 31 March 2004}% It is always \today, today,
% but any date may be explicitly specified
\begin{abstract}
We discuss a category of graphs, recursive clique trees,
which have small-world and scale-free properties and allow
a fine tuning of the clustering and the power-law exponent of their
discrete degree distribution.
We determine relevant characteristics of those graphs:
the diameter, degree distribution, and clustering parameter.
The graphs have also an interesting recursive property,
and generalize recent constructions with fixed degree
distributions.
\end{abstract}
\pacs{89.75.Hc, 89.75.Da, 89.20.Hh}
%89.20.Hh World Wide Web, Internet
%89.75.Da Systems obeying scaling laws
%89.75.Fb Structures and organization in complex systems
%89.75.Hc Networks and genealogical trees
%89.75.-k Complex systems
%05.10.-a Computational methods in statistical physics and nonlinear
% dynamics
\maketitle
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
In a complex system a large number of agents interact showing
a cooperative behavior which does not depend only on the
individual features of its parts, but also on its structure
and on the different sort of relations which can be established
between them and the environment. This global behavior allows
the system to attain certain achievements without the presence
of an administrative hierarchy or a central control mechanism.
A swarm of bees, an ant colony, companies which supply a
big city (water, electricity, telephone, etc.), even biological
mechanisms or social relationships are all complex systems where
global patterns emerge from the interaction of a large number
of similar elements.
In recent years, there has been an increase in understanding
of such complex systems in terms of networks, modeled by graphs
composed of vertices and edges, where vertices represent the
basic elements and edges their interactions.
The World Wide Web (WWW), Internet, transportation systems, and many
biological and social systems have been characterized by
small-world scale-free networks~\cite{WaSt98,BaAl99} which
have a strong local clustering (nodes have many mutual neighbors),
a small diameter (maximum distance between any two nodes) and
a distribution of the number of nearest neighbors for each
node according to a power law (scale-free property), see, for
example, Refs.~\cite{AdHu00,AlJeBa99,FaFaFa99,AlKuSt02}.
To model these networks and their growth very often
stochastic models and methods from statistical physics
have been considered; see Ref.~\cite{AlBa02}. However, the use
of exact deterministic models allows a quick determination
of the relevant parameters of the associated graph that may
be compared with experimental data from real
and simulated networks.
Previous work, for example, considered
deterministic small-world networks
comparable to those obtained stochastically
by Watts and Strogatz~\cite{CoOzPe00}, node replacement and node addition
methods to produce small-world and scale-free networks from
a low diameter ``backbone'' network~\cite{CoSa02},
specific recursive scale-free constructions with fixed degree
distributions~\cite{BaRaVi01,DoGoMe02,RaBa03,No03} and scale-free trees
(without clustering)~\cite{JuKiKa02}.
In this paper we present a recursive graph construction
which produces scale-free small-world networks with an
adjustable clustering and such that the parameter
of the power law
associated with the degree distribution, $\gamma$,
takes values between $2$ and $1+{\ln 3}/{\ln 2}=2.584 96$.
It has been shown that the values
for the scaling exponent of most real technical networks
~\cite{Me03} are in this range.
Moreover, the addition of one-degree nodes to our construction
allows a model which is not far from the real
topology of Internet at the autonomous system level, as is
described in Ref.~\cite{TaPaSiFa01}.
Therefore, we introduce here a deterministic exact network model
of complex systems for which we can adjust both
the clustering parameter and the power-law exponent,
providing a complementary tool
to stochastic approaches.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% DEFINITION of RECURSIVE CLIQUE_TREES
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Recursive clique tree $K(q,t)$, $q\geq 2$}. Definition.---
a complete graph $K_q$ (also referred
in the literature as $q$ clique; see Ref.~\cite{We01}) is
a graph without loops whose $q$ vertices are pairwise
adjacent.
The recursive clique tree $K(q,t)$ is the graph constructed
as follows:
For $t=0$, $K(q,0)$ is the complete graph $K_q$ (or $q$ clique).
For $t\geq 1$, $K(q,t)$ is obtained from $K(q,t-1)$ by adding
for each of its existing subgraphs isomorphic to a $q$-clique a
new vertex and joining it to all the vertices of
this subgraph (see Fig. \ref{fig:q3} for the case $q=3$).
Then, at $t=1$, $K(q,1)$ results in the complete
graph with $q+1$ vertices, $K_{q+1}$, and at $t=2$ we
add $q+1$ new vertices
each of them connected to all the
vertices of one of the $q$ cliques $K_q$ (subgraphs of $K_{q+1}$),
and so on.
This algorithm produces a complex growing
graph with a tunable parameter $q$
which controls all its relevant characteristics.
In the particular case $q=2$ we obtain the same
graph as in Ref. ~\cite{DoGoMe02}. However, our
family is infinite as $q$ can take any natural
value starting from 2.
Notice that although we call the graph a recursive clique tree,
the graph contains numerous cycles
and hence is not a tree in the strict sense.
Recursive clique-trees $K(q,t)$ are a natural generalization of trees
(if one considers the case $q=1$,
then we obtain a subfamily of the ``usual'' trees).
Similar constructions in which new vertices
are joined to every vertex of a given $q$ clique
have been considered, in another context,
in Ref. ~\cite{Bo88} and termed ``$k$ trees.''
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}
\includegraphics[width=65mm]{fig1.eps}% Here is how to import EPS art
\caption{\label{fig:q3} (Color online) First stages of a growing recursive clique tree $K(3,t)$. $K(3,0)$ is the complete
graph or clique $K_3$ (the triangle with thick edges). At each step a vertex is connected to each of
existing $K_3$. Therefore at $t=1$ only one vertex is added, but at $t=2$, four new vertices
should be connected to the four different cliques $K_3$.}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% RECURSIVE CONSTRUCTION OF CLIQUE_TREES
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Recursive construction of $K(q,t)$}.
There exists another interesting way to construct $K(q,t)$,
which clearly shows the recursive structure of such networks.
We call {\em native clique} of $K(q,t)$ its initial $q$ clique
at $t=0$. Then $K(q,t)$ is constructed as follows:
At step 0, we have the native clique $K_q$.
For any step $t\geq 1$, $K(q,t)$ is constructed as follows:
Consider a $(q+1)$ clique, then every subgraph of it
isomorphic to a $q$ clique (there are $q+1$ such cliques)
is a native clique of a $K(q,t-1)$; see Fig.~\ref{fig:recq2}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% FIG 2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}
\includegraphics[width=80mm]{fig2.eps}
\caption{\label{fig:recq2}(Color online) Recursive construction of $K(2,t)$.
We {\em glue} a native (or initial) clique of a $K(2, t-1)$ on each
$2$-clique of $K_3$ to construct $K(2,t)$. In this figure we
obtain $K(2,3)$ from three copies of $K(2,2)$ as follows: The graph at the left is $K(2,2)$ with
its native $2$-clique (the thick edge). Three copies of $K(2,2)$ are glued
by their natives $2$-cliques
to each of the three $2$-cliques of $K_3$ (center) resulting in $K(2,3)$ (right).}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% SIZE AND ORDER OF K(q,t)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Size and order of $K(q,t)$}.
Table~\ref{tab:table1} gives the number of new edges added
to the tree at each step and the total number of $K_q$
at this step.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% TABLE I
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{table}
\caption{\label{tab:table1} Number of new edges added to the recursive
clique tree $K(q,t)$ at each step $t$ and total
number of complete graphs $K_q$ at this step.}
\begin{ruledtabular}
\begin{tabular}{lll}
Step ($t$) & New edges & Number of $K_q$\\
\hline
0 & $\frac{q(q-1)}{2}$ & $1$ \\
1 & $q$ & $q+1$ \\
2 & $q(q+1)$ & $q(q+1)+(q+1)=(q+1)^2$ \\
3 & $q(q+1)^2$ & $q(q+1)^2+(q+1)^2=(q+1)^3$\\
$\cdots$ & $\cdots$ & $\cdots$ \\
i & $q(q+1)^{i-1}$ & $(q+1)^i$ \\
i+1 & $q(q+1)^{i}$ & $q(q+1)^i+(q+1)^i=(q+1)^{i+1}$ \\
$\cdots$ & $\cdots$ & $\cdots$ \\
\end{tabular}
\end{ruledtabular}
\end{table}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Therefore we can easily compute the total of edges
at step $t$,
\begin{equation}
|E|_t =\frac{q(q-1)}{2} + q\sum_{i=0}^{t-1}(q+1)^i=
\frac{q(q-1)}{2}+(q+1)^{t}-1
\end{equation}
The distribution of vertices and degrees is given in Table~\ref{tab:table2}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% TABLE II
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{table}
\caption{\label{tab:table2} Distribution of
vertices and degrees for the recursive
clique tree $K(q,t)$ at each step $t$.}
\begin{ruledtabular}
\begin{tabular}{lll}
Step ($t$) & Num. vertices & Degree \\
\hline
1 & $q+1$ & $q$ \\
%\hline
2 & $q+1$ & $2q$ \\
& $q+1$ & $q$ \\
%\hline
3 & $q+1$ & $q^2+2q$ \\
& $q+1$ & $2q$ \\
& $(q+1)^2$ & $q$ \\
%\hline
4 & $q+1$ & $q^3 +q^2+2q$ \\
& $q+1$ & $q^2+2q$ \\
& $(q+1)^2$ & $2q$ \\
& $(q+1)^3$ & $q$ \\
%\hline
$\cdots$ & $\cdots$ & $\cdots$ \\
%\hline
i & $q+1$ & $q^{i-1} +q^{i-2}+\cdots + q^2+2q$ \\
& $q+1$ & $q^{i-2}+\cdots + q^2+2q$ \\
& $\cdots$ & \\
& $(q+1)^{i-2}$ & $2q$ \\
& $(q+1)^{i-1}$ & $q$ \\
%\hline
$\cdots$ & $\cdots$ & $\cdots$ \\
\end{tabular}
\end{ruledtabular}
\end{table}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The maximum degree at step $i$ is
\begin{equation}
\Delta_i=\frac{q-q^i}{1-q}+q=\frac{q^i-1}{q-1}+q-1
\end{equation}
Adding up the number of vertices gives the result
\begin{equation}\label{Nt}
N_t=\sum_{j=1}^{t}(q+1)^j+(q+1)=\frac {(q+1 )^{t}-1}{q}+q
\end{equation}
The average degree is then
\begin{equation}
\overline{k}_t = \frac{2|E|_t}{N_t} =
\frac {q [ {q}^{2}-q+2\, \left( q+1 \right) ^{t}-2 ] }
{ \left( q+1 \right) ^{t}-1+{q}^{2}}
\end{equation}
[for $q=2$, it is $4/(1+3^{1-t})$]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% DEGREE DISTRIBUTION
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Degree distribution}.
The degree spectrum of the graph is discrete: at time $t$,
the number $N(k,t)$ of vertices of degree
$k = q, 2q, q^2+2q, q^3+q^2+2q, \ldots,
q^{t-2}+q^{t-3}+\cdots +q^2+2q, q^{t-1}+q^{t-2}+\cdots +q^2+2q$
is equal to
$(q+1)^{t-1}, (q+1)^{t-2}, (q+1)^{t-3},(q+1)^{t-4}, \ldots,
(q+1), (q+1)$,
respectively.
Other values of the degree are absent.
Clearly, for the large network, $N(k,t)$ decreases as
a power of $k$, so the network can be called ``scale free.''
Spaces between degrees of the spectrum grow with increasing $k$.
Therefore, to relate the exponent of this discrete degree
distribution to the standard
$\gamma$ exponent of a continuous degree distribution
for random scale free networks,
we use a cumulative distribution
$P_{cum}(k) \equiv \sum_{k^\prime \geq k}
N(k^\prime,t)/N_t \sim k^{1-\gamma}$.
Here $k$ and $k^\prime$ are points of the discrete
degree spectrum. For a degree
\begin{equation*}
k = q^{t-l}+q^{t-l-1}+\cdots +q+q
= q\left(\frac{q^{t-l}-1}{q-1}+1\right)
\end{equation*}
there are $(q+1)^{l-1}$ vertices
with this exact degree.
We count now how many vertices have this and a higher degree.
From the distribution
%%%%%%%%%
%\begin{eqnarray*}
%\lefteqn{\sum_{degree \geq k} = }\\
% &=& \sum_{p=l-1}^1 (q+1)^p + (q+1)\\
% &=& \sum_{p=1}^{l-1} (q+1)^p + (q+1)\\
% &=& \sum_{p=0}^{l-1} (q+1)^p + (q+1)-1\\
% &=& \frac{(q+1)^l-1}{q} +q
%\end{eqnarray*}
%%%%%%%%%
\begin{equation*}
\sum_{k' \geq k} N(k',t)= \sum_{p=1}^{l-1} (q+1)^p + (q+1) =
\frac{(q+1)^l-1}{q} +q
\end{equation*}
As the total number of vertices at step $t$ is given
in Eq.~(\ref{Nt}) we have
\begin{equation*}
\left[q(\frac{q^{t-l}-1}{q-1}+1)\right]^{1-\gamma} =
\frac{\frac{(q+1)^l-1}{q} +q}{ \frac{(q+1)^t-1}{q} +q }
= \frac{(q+1)^l-1+q^2}{(q+1)^t-1+q^2}
\end{equation*}
Therefore, for $t$ large
\begin{equation*}
(q^{t-l})^{1-\gamma}= (q+1)^{l-t}
\end{equation*}
and
\begin{equation}
\gamma \approx 1+\frac{\ln (q+1)}{\ln q}
\end{equation}
so that $2<\gamma <2.584 96$.
Also, notice that when $t$ gets large, the maximal
degree of a vertex is roughly
$q^{t-1} \sim N_t^{\ln q/\ln (q+1)} = N_t^{1/(\gamma-1)}$.
%
%which coincides with a standard relation for the cutoff of degree
%distribution in growing scale-free networks \cite{dm01c}.
%\smallskip
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% CLUSTERING
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Clustering distribution}.
The clustering coefficient $C(x)$ of a vertex $x$
is the ratio of the total number of existing
connections between all $k$ its nearest neighbors
and $k(k-1)/2$, the number of all possible connections between them.
The clustering of the graph is obtained averaging
over all its vertices. In what follows
${\cal N}_i(x)$ will denote the total number of neighbors
which $x$ has at step $i$, ${\cal K}_i(x)$ will denote the number
of different $q$ cliques which contain $x$ at step $i$, and
${\cal E}_i(x)$ will denote the total number of edges
among the neighbors of vertex $x$ also at step $i$.
\smallskip
Next we compute the clustering parameter for a vertex $x$ at any
step of iteration.
%\smallskip
{\em Step 1.}
As $x$ belongs to $K_{q+1}$, i.e., the recursive clique tree $K(q,t)$
at step 1, it is adjacent to $q$ other vertices [$ {\cal N}_1(x) =q$] which form a $K_q$.
Therefore the number
of edges among all vertices adjacent to $x$ is
${\cal E}_1(x)={q(q-1)}/{2}$.
Moreover, $x$ belongs to $q$ different $q$ cliques, i.e.,
${\cal K}_1(x)=q$.
%\smallskip
{\em Step 2.}
As the number of $q$ cliques in $K_{q+1}$ is $q+1$,
we add this number of new vertices to the construction.
But of these, $q$ will be forming --each-- a different
$(q+1)$ clique with $x$. Therefore
${\cal N}_2(x) =2q$ and ${\cal E}_2(x)={q(q-1)}/{2}+q(q-1)$.
Now $x$ will belong to ${\cal K}_2(x)={\cal K}_1(x)+q(q-1)=q^2$
different $q$ cliques.
%\smallskip
{\em Step 3.}
We add now to the construction ${\cal K}_2(x)$
new vertices which will form each a $(q+1)$ clique with $x$.
${\cal E}_3(x)={\cal E}_2(x)+ {\cal K}_2(x)(q-1)=
{q(q-1)}/{2}+q(q-1)+q^2(q-1)$.
${\cal K}_3(x)={\cal K}_2(x)+{\cal K}_2(x)(q-1)=q^2+q^2(q-1)=q^3$
and ${\cal N}_3(x)={(q^3-1)}/{(q-1)}+q-1$.
%\smallskip
{\em Step t.}
We add ${\cal K}_{t-1}(x)$ new vertices forming each
a $(q+1)$ clique with $x$.
${\cal E}_{t}(x) ={\cal E}_{t-1}(x) + {\cal K}_{t-1}(x)(q-1)=
{q(q-1)}/{2}+(q-1)\sum_{i=1}^{t-1}q^i=
(q-1)[{q}/{2}+{q^t-1}/{(q-1)}-1]$.
${\cal K}_{t}(x)={\cal K}_{t-1}(x)+{\cal K}_{t-1}(x)(q-1)=
{\cal K}_{t-1}(x)q=q^{t}$
and ${\cal N}_t(x)={(q^t-1)}/{(q-1)}+q-1$.
%\smallskip
Therefore the clustering for vertex $x$ after $t$ iterations is
\begin{equation}
C(x)=\frac{2{\cal E}_{t}(x) }{ {\cal N}_t(x) [ {\cal N}_t(x)-1]}
\end{equation}
Notice that $ {\cal N}_t(x) = \Delta_t$ is precisely the maximum degree
at step $t$.
We see easily that there is a one-to-one correspondence
between the degree of a vertex and its clustering.
In general, we will have that for a vertex $v$ of degree $k$
the clustering parameter will be
\begin{equation}
C(v)=\frac{2(q-1)(k-\frac{q}{2})}{k(k-1)}
\end{equation}
Notice that for $q=2$ we have
$C(v)=2/k$ as in Ref. \cite{DoGoMe02}.\\
Using this result, we can compute now the clustering of the graph,
%%%%%%%%%
%\begin{eqnarray*}
%\lefteqn{\overline{C}_t =\frac{2(q+1)(q-1)(\Delta_t-\frac{q}{2})}{\Delta_t(\Delta_t-1)}}\\
% &+& \sum_{i=1}^{t-1} \frac{2(q+1)^{t-i}(q-1)(\Delta_i-\frac{q}{2})}{\Delta_i(\Delta_i-1)}
%\end{eqnarray*}
\begin{equation}
\small
\overline{C}_t =\frac{2(q+1)(q-1)(\Delta_t-\frac{q}{2})}{\Delta_t(\Delta_t-1)}
+ \sum_{i=1}^{t-1} \frac{2(q+1)^{t-i}(q-1)(\Delta_i-\frac{q}{2})}{\Delta_i(\Delta_i-1)}
\end{equation}
For $t\geq 7 $ and for any $q\geq 3$ we have
\begin{equation}
\overline{C}_t\geq\frac{3q-2}{3q-1}
\end{equation}
Thus the clustering is high, and, similarly to the $\gamma$
coefficient of the power law, is tunable by choosing
the right value of $q$: in particular, $\overline{C}$ ranges
from $\frac{4}{5}$ (in the case $q=2$) to a limit of 1 when $q$ gets large.
%\smallskip
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\em Diameter}.
We recall that the diameter of a graph is the maximum of all
distances between any two vertices of the graph.
%Given any two nodes $u$ and $v$ in $K(q,t)$, the {\em distance} between
%$u$ and $v$ is the smallest number of jumps that are necessary to
%reach $v$ from $u$. The {\em diameter} of $K(q,t)$ is the longest
%distance that exists among pairs of nodes of $K(q,t)$.\\
Computing the exact diameter of $K(q,t)$ can be done analytically, and
gives the result shown below. Here, we present the main ideas of
this analysis.
First, we note that for any step $t$, the diameter always lies between
pair of vertices that have just been created at this step.
We will call such vertices {\em outervertices}. Then, we note
that, by construction, no two outervertices created at the same step
$t$ can be connected. Hence, if $v_t$ is an outervertex created at step $t$,
then $v_t$ has been connected to a $q$ clique composed of vertices
created at pairwise different steps $t_1<t_2<\cdots<t_q$ and consequently, $t_1\leq t-q$.
Now we want to know the distance between two outervertices $u_t$ and
$v_t$. The idea is, starting from $u_t$, to ``get back'' to $K(q,1)$
by jumps from $u_t$ to $u_{t-q}$, then to
$u_{t-2q}$, etc. Hence it takes at most about ${t}/{q}$ jumps to go from
$u_t$ to $K(q,1)$. Then it takes at most as many jumps to go
from $K(q,1)$ to $v_t$ and we conclude that the diameter
cannot be bigger than (roughly) ${2t}/{q}$.
More precisely, the exact formula for the diameter of $K(q,t)$,
denoted $Diam[K(q,t)]$, is the following~:
\begin{equation}
Diam[K(q,t)]=2\left(\lfloor\frac{t-2}{q}\rfloor +1\right)+f(q,t)
\end{equation}
where $f(q,t)=0$ if $t-\lfloor{(t-2)}/{q}\rfloor q\leq \lceil{(q+1)}/{2}\rceil$, and 1 otherwise.
This value can be obtained by a sharp analytical proof.
When $t$ gets large, then $Diam[K(q,t)]\sim {2t}/{q}$, while
$N_t\sim q^{t-1}$, thus the diameter clearly grows logarithmically
with the number of vertices.
{\em Discussion}.
The recursive clique-trees $K(q,t)$ which we study in this paper are
recursively constructed graphs which have both
small-world and scale-free characteristics.
By choosing adequately the value of $q$
it is possible to obtain different
clustering parameters and power-law exponents
(the clustering of the graph ranges from $0.8$ to $1$
while the power-law exponent
takes values between $2$ and $2.584 96$).
It has been shown that many networks modeling complex systems such as
the WWW, Internet, movie actors, the Erd\H{o}s number, interactions in protein complexes
of {\em Saccharomyces cerevisae}, the metabolic network of {\em Escherichia coli}, etc.,
have clustering and power-law parameters precisely in these ranges, see, for
example,
Refs. \cite{BaDeRaYoOl03,FaFaFa99,WaSt98,Me03,JeToAlOlBa00,GoOhJeKaKi02}.
Moreover, the recursive clique trees are actually a
deterministic tunable generalization of the
scale-free growing networks introduced in Ref.~\cite{DoMeSa01} in which
one vertex is created per unit time and connects
to both the ends of a randomly chosen edge. These networks were further
generalized and called ``pseudofractal'' graph
in \cite{DoGoMe02}, and correspond to
the particular case $q=2$ of the recursive clique trees.
Variations of our construction, for example, by choosing
that in different steps the new vertices added are attached
to cliques of different size, would allow a
richer structure and more flexibility in the
control of the clustering and power-law exponent and other relevant
network parameters.
\smallskip
The authors thank a Picasso bilateral grant, references
HF2001-0056 (Spain) and 04295SG (France).
Additional support for the first author was
provided by the Ministry of Science and Technology, Spain, and the
European Regional Development Fund (ERDF) under Project No.
TIC2002-00155.
\begin{references}
\bibitem{BaAl99} A.-L. Barab\'asi and R. Albert,
% Emergence of scaling in random networks,
Science {\bf 286}, 509 (1999). %pp. 509--512
\bibitem{WaSt98} D.J. Watts and H. Strogatz,
% Collective dynamics of `small-world' networks,
Nature (London) {\bf 393}, 440 (1998). %pp. 440--442
\bibitem{AdHu00} L. Adamic and B. Huberman,
% Power-law distribution of the world wide web,
Science {\bf 287}, 2115 (2000).
\bibitem{AlJeBa99} R. Albert, H. Jeong, and A.-L. Barab\'asi,
% Diameter of the world wide web,
Nature (London) {\bf 401}, 130 (1999). %pp. 130--131
\bibitem{AlKuSt02} E. Almaas, R.V. Kulkarni, and D. Stroud,
% Characterizing the structure of small-world networks,
Phys. Rev. Lett. {\bf 88}, 098101 (2002).
% http://link.aps.org/abstract/PRL/v88/e098101
\bibitem{FaFaFa99} M. Faloutsos, P. Faloutsos, and C. Faloutsos,
% On power-law relationships of the internet topology,
Comput. Commun. Rev. {\bf 29}, 251 (1999).
\bibitem{AlBa02} R. Albert and A.-L. Barab\'asi,
% Statistical mechanics of complex networks,
Rev. Mod. Phys. {\bf 74}, 47 (2002). %pp. 47--97
\bibitem{CoOzPe00} F. Comellas, J. Oz\'on, and J.G. Peters,
% Deterministic small-world communication networks,
Inf. Process. Lett. {\bf 76}, 83 (2000). %pp. 83--90
\bibitem{CoSa02} F. Comellas, and M. Sampels,
% Deterministic small-world networks,
Physica A {\bf 309}, 231 (2002). %pp. 231--235
%ISSN 0378-4371 .
\bibitem{BaRaVi01} A.-L. Barab\'asi, E. Ravasz, and T. Vicsek,
% Deterministic scale-free networks,
Physica A {\bf 299}, 559 (2001). %pp. 559--564.
\bibitem{DoGoMe02} S.N. Dorogovtsev, A.V. Goltsev, and J.F.F. Mendes,
% Pseudofractal scale-free web,
Phys. Rev. E {\bf 65}, 066122 (2002).
\bibitem{RaBa03} E. Ravasz and A.-L. Barab\'asi,
% Hierarchical organization in complex networks.
Phys. Rev. E {\bf 67}, 026112 (2003).
\bibitem{No03} J.D. Noh,
% Exact scaling propertiesof a hierarchical network model.
Phys. Rev. E {\bf 67}, 045103 (2003).
\bibitem{JuKiKa02} S. Jung, S. Kim, and B. Kahng,
% Geometric fractal growth model for scale-free networks.
Phys. Rev. E {\bf 65}, 056101 (2002).
\bibitem{Me03} J.F.F. Mendes, in {\em Statistical Mechanics of Complex Networks},
edited by R. Pastor-Satorras, M. Rubi, and A. Diaz-Guilera,
Lecture Notes in Physics Vol. 625 (Springer, Berlin, 2003), pp. 88--113.
\bibitem{TaPaSiFa01} S.L. Tauro, C. Palmer, G. Siganos, and M. Faloutsos,
% A simple conceptual model for the Internet topology
in {\em Global Telecommunications Conference GLOBECOM '01} % 25-29 Nov. 2001
(IEEE, Piscataway, NJ, 2001), Vol 3, pp. 1667 -1671.
\bibitem{We01} D.B. West,
{\em Introduction to Graph Theory}
(Prentice-Hall, Upper Saddle River, NJ, 2001).
% ISBN 0-13-014400-2
\bibitem{Bo88} H.L. Bodlaender,
% Some classes of graphs with bounded treewidth,
Bull. EATCS {\bf 36}, 116 (1988). %pp. 116--126
\bibitem{JeToAlOlBa00} H. Jeong, B. Tombor, R. Albert, Z. N. Oltvai, and A.-L. Barab\'asi,
% The large-scale organization of metabolic networks ,
Nature (London) {\bf 407}, 651 (2000). %pp. 651--654
\bibitem{GoOhJeKaKi02} K.I. Goh, E.S. Oh, H. Jeong, B. Kahng, and D. Kim,
% Classification of scale free network,
Proc. Natl. Acad. Sci. U.S.A. {\bf 99}, 12583 (2002). % 12583--12588
\bibitem{DoMeSa01} S.N. Dorogovtsev, J.F.F. Mendes, and A.N. Samukhin,
% Size-dependent degree distribution of a scale-free growing network,
Phys. Rev. E {\bf 63}, 062101 (2001).
% URL={http://link.aps.org/abstract/PRE/v63/e062101},
\end{references}
\end{document}
%
% ****** End of file ******