CELLULAR AUTOMATA (CA): A BRIEF

Generally, nonlinear relations among elements of complex systems and feedback loop give birth to unstable unforecastabilities whether the scale. In order to follow up “appearances” within it, the urban system should be analyzed from the smallest element to the whole and the model should absolutely include every scale for a correct CA design. As proved in physics and in chaos theory, basic structural elements play an important role in determining general characteristics of a system as a whole (Haken; 2006:6). Monitoring an urban system just as a whole may be illusive since it does not include or ignore what happens at the scale of its basic elements. Behavioural

basic elements and buildings and/or land plots are highly suitable to understand an urban dynamic as a system (Akbulut; 2004; Akbulut, Başlık; 2011). Hence, buildings or land plots should be considered as the smallest urban element like a cell of an organism. According to Batty’s definition the smallest unit of a city is the “cell” for physical space and “agents” (humans) for social units. The smallest units display a transformation from bottom to the top or, from local to universal in other words. The behaviour pattern in local level of the smallest unit, defined as agent will result of appearances in global scale. Cellular Automata is one of convenient models for this structure. In fact, CA is also an agent based model since cells play the role of agents (Batty; 2009:57). Cellular Automata (CA), at this point, helps us to understand dynamic behaviours of complex systems and predict future actions since it is an appropriate tool for behavioural fundaments of complex systems. As a powerful model able to measure spatial transformation due to influential grand projects and investments on urban structure, Cellular Automata, provides an important contribution to urban and regional planning for observing and understanding behaviours of complex urban systems, to predict urban growth. However, CA is also criticized for tending to be indicative rather than predictive particularly in predicting urban growth (Batty; 2009:57).

Although, CA was in use in analysis of complex systems in many disciplines since 1970s, it was only introduced a decade later in 1980 in urban and regional planning. It is claimed that this delay is mostly due to retarded awareness of geographers and urban planners in perception, comprehension, research and analysis of human agglomerations and regional structures within the concept of systems theory (Benenson; Torrens; 2004:72). CA was also unfamiliar at the beginning to geographers and urban planners’ conventional spatial

analysis methods on the spatial base of defined regions or territories (zones). Whereas, the “cell” is the base unit of CA in modelling whether a life cycle or a city. Use of CA in urban and regional researches is particularly gained a common ground following spread of geographical information systems (GIS) and adoption of the method of definition of datas in grid pattern of cells. In general use, territories determined by administrative boundries or service areas is replaced by pixels and grid attributed datas (Batty; 2007:16). Raster analysis techniques, remote sensing images and digital land use maps in GIS provides a natural base for CA. Even, CA modules in some GIS software highly simplified analysis of land use transformation, future predictions and urban simulations.

Human agglomerations are generally considered, conceptualized and associated with different forms. The modernist approach of city as a machine is maybe the most familiar one. Another common conceptualization of city is to consider it as an organism. Emergence, namely the birth, expansion, the grow up, transformations, the maturation or mutation and even abandonment and disappear namely, death all display an organism’s life cycles in a city’s life and dynamics or different phases, a city undergoes may be associated to life cycles of an organism. Spatial sprawl of a city may then be interpreted as proliferation of cells, growth in numbers. This is obviously not for the sake of a pleasant mind game or just to make some similarities. Thinking of a city like an organism is a groundbreaking contribution to our knowledge base and understanding about cities. Contribution and innovative approach to some resident urban problems such as prediction of urban growth and spatial transformation is maybe the primary benefit from CA.

The matrix of cells, each of which cover an equal area such as an acre, a hectare, a square-kilometer, a square-meter etc. is the main essence of CA and through observation and analysis of behaviour of cells like an organism, spatial growth, urban sprawl, macroform, functional and land use change and transformation of a city become more predictable. In addition to matrix of cells, a CA model, only contains a few parameters namely, a state, input, transition rule, neighbour and time. A state is the probability of what a cell may be in a condition of two options only like open-closed, 1-0, action-inertia, road-railroad or it may refer to an information of situation among land use alternatives like trade, housing, industry etc. Any cell organizes itself according to state of neighbouring cells and the input is the information a cell receives from outside. Rules are determinant sentences to define a behaviour pattern according to a received information from outside. Rules are in accordance with the time of a cell in a given situation and define the next state of a cell according to received information from outside. Neighbouring is one of most important parameters of a CA. Definition of neighbouring cells may vary according to aim, spatial characteristics and time.

Transition rule is another important factor in working principles of a Cellular Automata and this make it works and effective. Here, rules may be conditional statements with “if, then” declarations and can contain mathematical expressions and functions. Transition rules should be applicable to each cell, in every situation and all times (Torrens; 2000).

Working principles of Cellular Automata may look simple but, many different situations should be considered in complex urban systems according to every single local characteristics. CA space of cell or working unit in other words represents inhomogeneous pieces of a

city since every pieces of a city which corresponds to cells has also different spatial and physical particularities such as topography, land use, accessibility, distance to a center etc.

4. THE METHOD: CELLULAR AUTOMATA AND LAND