Information Technology Reference
In-Depth Information
Figure 7.2.
The equilibrium positions as a function of the coupling parameter K , redrawn from [ 15 ] with
permission.
By inserting the definition of the transition rates given by ( 7.16 )into( 7.18 ) we obtain
the mean-field equation for the web of two-state nodes,
d
dt =−
) =−
U
()
,
2 g
cosh
(
K
)
2 g sinh
(
K
(7.19)
which corresponds to the dynamics of a particle in a double-well potential introduced
in Chapter 3.
Equation ( 7.19 ) describes the overdamped motion of a particle, whose “position” is
, and the values of its minima depend only on
the coupling constant K . It is straightforward to show that there is a critical value of the
coupling parameter given by K
within the symmetric potential U
()
=
K c =
1
,
such that (1) if K
K c the potential has
only one minimum located at
=
0; or (2) if K
>
K c the potential is symmetric and
has two minima located at
± min separated by a barrier with the maximum centered at
=
The potential has the same form as the quartic potential shown in Figure 3.4 and
the height of the potential barrier is a monotonic function of the coupling parameter K .
The time evolution of the dynamical variable
0
.
is determined by the extrema of
the potential and consequently two kinds of dynamical evolution are possible: (1) if
K
K c , the dynamical variable
(
t
)
will, after a brief transient, settle to an asymp-
totic value
( ) =
0 independently of the initial condition
(
0
) ;
(2) if K
>
K c ,the
dynamical variable
(
t
)
will, after a transient, reach the asymptotic value
( ) =
min =
0 for an initial condition
(
0
)>
0
,( ) =− min =
0 for an initial condi-
tion
(
0
)<
0 and finally
(
t
) =
0 for all time given an initial condition
(
0
) =
0. In
Figure 7.2 we compare the minima
± min and the numerical evaluation of
( )
for
various values of the coupling constant K .
The location of the potential minima as a function of the coupling parameter K is
given in Figure 7.2 , which shows that a phase transition occurs at K
=
K c =
1. For
a single two-state node the condition
( ) min =
p 1
p 2 =
0 corresponds to
the statistical preference of the particle to be in either the state r
2.
This is a consequence of the transition rates being different if the coupling parameter
is greater than the critical value. By inserting the mean-field value of the probabilities
( 7.15 ) into the expression for the transition rates ( 7.16 ) and allowing the dynamical
variable to reach its asymptotic value we obtain
=
1 or the state r
=
ge K ( ) =
ge K ( ) .
g 12 =
g 21 =
(7.20)
 
Search WWH ::




Custom Search