Cellular Automata Simulation 187
(a) α = 1.0, no signal (b) α = 0.8, no signal (c) α = 0.8, signal type 110
FIGURE 7.5: Congestion simulation using BCA.
An in-depth explanation of CAs describing traffic congestio n is given in section
7.9.
Figure 7.5 shows a result of BCA simulation. Cars are shown in black
in this simulator. Continuous black areas represent congestion because cars
cannot move fo rward. Congestion is more likely to form when the probability
that a car moves forward α is small. Figure 7.5(a) shows how congestion forms
(black bands wider than two lattice points form where a car leaves the front
of the congestion and another car joins at the rear). A signal is added at the
center in Figure 7.5(c). The signal pattern is shown in blue when 1 and red
when 0, and the pattern in this simulation was 110 (i.e., blue, blue, red, blue,
blue, red,. . .).
The Swarm simulator using BCA is an applicatio n of one-dimensional CA,
and the par ameter probe in ObserverSwarm displays the fo llowing parameters.
• worldSizeX: size of the spac e (horizontal a xis)
• worldSizeY: size of the spac e (vertical axis)
• Alpha: pro bability that a ca r moves to the right
• Signal1: pattern of signal 1
• Signal2: pattern of signal 2
The proba bility that cars move to the right and the pattern of signals can
be changed during runs by clicking the “Stop” button, changing values, and
re-running. Here, after changing the parameters, you need to press “Enter”
and click the applyRule button.
A s imulation of silicon traffic based on these models is shown in Fig. 7.6.
A two-dimensional map is randomly generated, and cars are also placed at
random. Two points are randomly chosen from the nodes on the map and
are designated as the or igin and the des tina tion. The path from the orig in to
the destination is determined by Dijkstra’s algorithm. Here, a co st calculation
where the distances between nodes are we ighted with the density of cars in
each edge is performed to determine a path that avoids congestion. The signal