We quickly review the one-lane CA model of
[1]. In simulation of this and other
CA models to be discussed below, it is convenient to apply
the rule to a lattice with periodic boundary conditions.
Each site may be
empty or occupied by a car with an integer velocity
.
gives good agreement with physical experiments.
The variable 
gives the number of unoccupied sites in front of a vehicle.
is the probability to randomly decelerate, and
is a random number between 0 and 1.
One iteration consists of the three following sequential steps, which are
applied in parallel to all cars:
THEN 
THEN 
AND 
rand 
) THEN 
These simple conditions already give realistic results. Step 3 accounts for stochastic driver behavior.
The two-lane uni-directional model is based on two
parallel single-lane models [2].
Four additional conditions account for the exchange of
vehicles between the lanes.
First the vehicles change lanes, then the
one-lane algorithm is applied.
This model introduces several new variables:
: the number of unoccupied sites
in front of vehicle 
on the (same, opposite) lane, respectively.
the number of unoccupied sites
behind the vehicle, on the opposite lane;
, 
: minimum free distance needed for a pass,
ahead on same lane, ahead and behind on the opposite lane;
: the probability to change lanes. The added rules are as following:
) AND
) AND
) AND
) THEN change lane
This model has a symmetric and an asymmetric version. In the asymmetric version, one is not allowed to pass on the right.