GeoSOS

</noinclude>GeoSOS is the abbreviation of a computer-based system: Geographical Simulation and Optimization System . It integrates three components: Cellular Automata (CA), Multi-Agent System (MAS) and Swarm Intelligence (SI). The functions of GeoSOS include simulation, prediction, optimization and displaying geographical patterns and processes (http://www.geosimulation.cn). Since the conventional geographical information systems (GIS) have a general lack of ability to address spatial process, GeoSOS can be regarded as a complement for current GIS.
Applications
Currently, applications of GeoSOS are mainly in the cities of the Pearl River Delta, China, such as Guangzhou, Panyu, Dongguan , and Shenzhen . There are other applications in China and other countries which are not published in papers.
Feature
Cellular Automata
Five models are provided in the CA simulation component, including MCE-CA, Logistic-CA, PCA-CA, ANN-CA, and Decision Tree-CA(Li and Yeh 2004). In MCE-CA, a multi-criteria evaluation based on analytical hierarchy process is used to define the weights of involving spatial factors. Development suitability is then calculated and further transformed into the development probability. Logistic-CA is the updated version of MCE-CA by using logistic regression . While in PCA-CA, a principle component analysis is implemented to reduce the correlation among factors and transform them into several components. The ANN-CA uses artificial neural networks to determine transition rules and is capable of simulating changes of multiple land use types. The Decision Tree-CA integrates a decision tree algorithm in urban and land use simulation. It uses explicit rules in a “IF…THEN…” manner which provides better understandings of the simulation than the manners of using coefficients or equations.
Multi-Agent Systems
The component for multi-agent simulation now contains a demo about land price simulation using a multi-agent systems method. It supposed to be a modeling environment for multi-agent systems simulation, allow users to build a multi-agent model by a GUI, or by a embed language in the GeoSOS platform, also can integrate a CA method with the CA module. This module is still under construction.
Spatial Optimization
The spatial optimization component contains three models for site selection, path optimization and land allocation respectively. These models are all based on algorithms of ant colony optimization . The site selection model is to find the best locations of facilities by maximizing the benefits and minimizing the costs . In path optimization, the objective is to maximize the coverage while maintaining the length as short as possible . In the land allocation model, a land use suitability map in raster format is required, and given the amount of land to allocate, the model generates an allocation alternative that considering maximization of both land use suitability and compactness of resulted patterns.
Architecture
GeoSOS is implemented by using Microsoft .NET Framework and C# programming language. The software is designed to be a plug-in architecture, which facilitate the different methods and algorithms of cellular automata, multi-agent systems and spatial optimization as a plug-in, then plug to or pull out the whole GeoSOS framework.
The whole GeoSOS software suite includes a standalone application, some ESRI ArcGIS add-ins, and some .NET dll class libraries. Also GeoSOS has two virtual globe environments to display the simulation outcomes, one is an application which is developed in NASA World Wind, another is a web plug-in of World Wind which can be browsed in web pages. Now the standalone application is available in the GeoSOS website.
Design
One novelty of GeoSOS is its capability of coupling simulation models with optimization models. More effective simulation and optimization tasks can be accomplished under this coupling framework .
Another novelty of GeoSOS is the integration of CA with MAS, which allows the system to address some social and human factors for handling more complex simulation tasks.

Main Menu

See Also

Main Menu