Cellular Automata Based Temporal Process Understanding of Urban Growth - Cellular Automata - page 334

three sets of weight values (Table 1). In other words, the temporal process can enable
us better understand the organized local growth. If we explore the changes of weight
values, it can be found that the major changes are indicated in major road and center.
As explained in section 2 (eq. 3 and 4), weight values should be the functions of tem-
poral development demand. Table 1 also shows the functions are highly complicated
in reality. A universal or standard function is not available. Rather it should be simpli-
fied and based on local knowledge. Model 2 actually is based on the interviews with
local planners.

Table 1. Test of temporal heterogeneity (Zuankou)

Results Model 1 Model 2

Total cells 1390 1390

Accuracy (CC) 55% 55%

Lee-Sallee Index 38% 38%

Stochastic
(α=10%)

1% 1%

Iteration number 50 50

Neighb. size 6 6

c 0.5 0.5

Temporal division 100% <15%, 15%-50%, >50%

Major road (MR) 0.2 - 0.5 0.05

Minor road (OR) 0.3 - 0.1 0.15

Centers (CE) - 0.7 - 0.5

Neighb. (new) 0.3 0.3 0.1 0.15

Master Planning 0.2 - 0.3 0.15

Total 100% 100% 100% 100%

Gradient for MR 0.0005 0.0005

Gradient for OR 0.002 0.002

Gradient for CE - 0.0004

Note( "-" :0)

Fig. 2 . Test of temporal heterogeneity (Model 1 & Model 2)