by : 1 Hu Shiyuan and 2. Li Deren
1. School of Resource and Environmental Science , Wuhan University, 129 Luoyu Road, Wuhan, China, 430079
2. School of Remote Sensing Information Engineering, Wuhan University, 129 Luoyu Road, Wuhan, China, 430079
Cellular automata (CA) are simple mathematical systems that exhibit very complicated
behaviour. The integration of GIS and CA shows tremendous capability in simulating
spatio-temporal dynamic process in geography world. But standard CA has some
restrictions in cellular shape, spatial resolution, precision, quantity, neighbor and
rule, which restrict the CA’s abilities of simulating real world and other applications.
This paper discusses the relation between geography phenomenon and cellular automata, extends the constitutes of standard CA such as cell and states, neighborhoods, local transition rule and discrete time and so on, and builds a extended CA model based on geographical entity in irregular geographical spaces, where cells are no longer regular tessellations, and cellular neighborhoods vary from place to place, and are no longer based on physically adjacent cells. In addition, a general integrated framework of GIS and vector cellular automata based geographical entity (VCA) is brought forward in the paper.
Cellular automata (CA) are mathematical models for systems in which many simple
components act together to produce complicated patterns of behavior (Wolfram, 1985).CA
have close associations with complexity theory and have been employed in the exploration
of a diverse range of urban phenomena. Urban applications of CA range from traffic
simulation and regional-scale urbanization to land-use dynamics, historical urbanization,
and urban development. The integration of GIS and CA will accelerate GIS’s ability of
simulating geographical process greatly especially (Zhou et al., 2001).
CA models are usually based on fine regular tessellations such as a grid, in which every cell
is identical, has identical relations with each of its neighbors, and has a standard
neighborhood of cells in which these relations operate. These neighborhoods are strictly
local in that they are based on physically adjacent cells. In geographic and urban models,
this may be over-simplistic (O'Sullivan, 2000), and it has some restrictions in cellular
shape, neighborhood and neighbor rules,which restrict the CA’s ability to simulate real
world.The standard CA exists some problems mainly as follows: (1) Space partition,
namely determination of space pixel. Each kind of graphical object has itself space scale in
the system which plenty of graphical entities exist together. In addition, graphical entity
represents different behavior in different space scale. It is a problem how to determinate a
uniform spatial resolution. (2) Precision & Quantity. CA models are usually based on fine
regular tessellations, cell is similar to the grid of grid data in GIS, it exists some problems
such as imprecise locating and tremendous quantity. (3) Cell space is divided into regular
tessellations on abstract space in standard CA. Every cell is identical, has identical relations
with each of its neighbors. This kind of CA can expose local reciprocity among cells. But geographical system is a typical complex system, which is a compound system consisted of
physical, social and economic subsystems. The complexity is an essential characteristic of
Geo-Spatial System for its complexity properties such as non-equilibrium, multi-scale,
indeterminacy, hierarchy, self-organizing, self-similarity, randomicity, iterativeness, and so
forth. So regular space system exists hardly in real world.
This paper will explore the relation between geographical space and cellular automata,and
build a extended CA model based on geographical entity in irregular geographical spaces.
In addition, this paper will explore the integrated pattern of GIS and CA.