V (b) is taken as state space of a Markov chain. Two requirements:
the limiting probability must be uniform and must be reached quickly.
Markov chain: Given a solution x = {x1,
x2,..., xn),
the transition to the next state y is defined as follows: with
probability 1 - a (1 > a
> 0) let y := x ,otherwise select randomly
and uniformly j(1 £ j £
n); let x' := {x1,
x2,..., xn}
with
xi = xi
i ¹ j, xj=
1 - xj; if ax £
b then y := x else
y :=x.
