THE PRINCIPLES OF GEOGRAPHICAL EXPERT SYSTEM 
(Professor) Ma Ainai 
(Doctor Candicate) Zhou Zhangfa 
(Geography Department, Peking University, China) 
(ISPRS Commission III) 
ABSTRACT 
AS computer technologies used in geography, Geographical Information System (GIS) is developed 
to manage geographical data and to solve some geographical problems according to geographical 
models.But there are many geographical problems that can not be defined by definite models and 
can not be solved by GIS. As the combination of geography and Artifical Intelligence (AI) 
, 
Geographical Expert System (GES)is designed to deal with these kinds of problems.GES is a kind 
of software system, which is based on geographical knowledge from geographer to deduce 
intelligent results.Therefore GES can deal with many geographical problems that often can only 
be solved by experts. 
KEY WORDS: Geographical Expert System, Geographical Information System, Geographical Knowledge 
Base, Geographical Inference Engine. 
1. INTRODUCTION 
Geography as a science has been on the phase 
of concentrating its research mainly on 
describing phenomena for a Tong time. 
Geographer based on predecessors working 
results and geographical data and 
descriptions got from field trip to draw a 
serials of conclusion, and those conclusion 
will become experience for later geographer. 
Figure 1 shows the classical geographical 
methods. 
  
When to process these kinds problems, 
geographer use geographical knowledge and 
experience to reason and gain inference 
conclusion. Expert System (ES) use knowledge 
as reasoning kernel to imitate human expert 
reasoning and get proper results. When we use 
ES technology in geography, we can build 
Geographical Expert System (GES) , which is a 
Kind of computer system with a great deal of 
geographical knowledge, and can reason with 
  
  
  
  
  
  
  
  
  
  
Previous 
work ing 
results |——| Data (map, 
report,...) 
arrangement  |——— 
Field —— | & processing 
work ing 
(trip & 
point) 
Fig. 1 
By the end of 1960s, with the development of 
Remote Sensing and computer technology, 
geography has adopted many new research 
methods, which is changing geography from 
qualitative science to quantative one.Now the 
research methods of geography has reached a 
new level, especially in Geographical 
Information System (GIS) . GIS not only gains 
and manages geographical data from different 
sources, such as remote sensing images and 
field measured data,but also derives some new 
information by calculating geographical data 
according to geograhical models. GIS now is 
successfully used for geographical planning 
and resource management, such as landuse plan 
urban developement. On the other hand,most of 
geographical problems are inexact, which can 
not be or are very difficult expressed as 
definite models, for example, how to select 
suitable methods for the conservancy of soil 
and water, make an auxiliary decision of 
urban system, make an agricultural planning, 
etc. As a kind of Management Information 
System (MIS), GIS is not suitable to process 
these kinds of problem. 
  
  
  
New concepts conc lus ion 
and — | (map, report, 
models table, ...) 
  
  
  
  
  
269 
  
Classical geographical methods 
these knowledge to solve inexact and 
undefinite geographical problems. 
GES 1s the combination of geography and 
Artifical Intelligence (AI) , and can be 
def inated as 
GES = GIS + artifical intelligent kernel. 
Supported by Remote Sensing, GIS and GES, 
Modern geographical methods can be shown as 
figure 2. 
2. DESIGN OF GES 
Designed to solve geographical problems by 
reasonning with geographical knowledge, GES 
put geographical knowledge at the most 
important position. As Geographical data is 
the inference source, it is the reasonning 
object with geographical knowledge. There are 
three main problems in GES, which are 
concerned with geographical knowledge and 
data,that is,