IN
only
place
omic
ns in
) this
xpert
es in
data
with
ar or
1odel
1 and
on to
ch is
stem
ty of
nter-
data
t tool
IS to
patial
pport
ctice.
used
ment
ilizer
much
| uses
those
ected
lems.
lexity
)CESS,
)CESS.
ledge
ring.
> and
nese
lacks
very
jlture
edge-
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
based and GIS integrated system as a public service tool will be
a great step to enhance agriculture modernization in China.
We will use GIS as a basic tool to store, extract and display
spatial data, while domain models and special knowledge are
integrated and collaborated to give a concrete decision result in
agriculture field management. Our objective is to address the
problem of field variability with location and plant variation to
give strategies of water and fertilizer application aiming to
enhance production or reduce invests input in agriculture. The
approach we developed for this purpose involves acquiring
sufficient expert knowledge, building models and designing
casy-to-understand result presentation to support farmers with
executable field management decision. GIS and Knowledge-
based decision support system, GZAgri-GIS is such an
information system that has been used in practice in Guangzhou,
China, aiding farm-level agriculture decision-making. We
integrate decision knowledge, decision models and GIS basic
functions in the system design and implementation. Since the
system intends to server the public, the framework of
Browser/Server is adopted.
2. BACKGROUND
2.1 Overview of SDSS
Expert system or ES is a computer program that reasons with
the knowledge of a specialist subject with a view to solving
problems or giving advice. Such a system may completely
fulfill a function that normally requires human expertise, or it
may play the role of an assistant to a human decision-maker. As
some functions can be performed by domain models, models
are then be used together with expert knowledge to get a more
secure result. This expert system with model component is the
preexistence of DSS, with high capability in numerical
calculation. DSS were first introduced in business management,
commercial investment and activities as strategic planning,
scheduling of business operations, and investment appraisals.
The first DSS applications began to appear in the early 1970s.
Since the early 1980s, DSS developed much under influence of
the PC revolution, the increasing performance price ratio of
hardware and software. Although there is not a generally agreed
upon definition, the term DSS commonly refers to "computer-
based systems which help decision makers utilize data and
models to solve ill-structured problems". In recent years, spatial
problems such as site location selection, shortest route selection,
and resource distribution plan are easy to see in our life. The
definition of SDSS can take many forms. But it is generally
agreed that SDSS is evolved from DSS and defined as a DSS
which combines geographic information with appropriate
algorithms and extend these capabilities to provide a rational
and objective approach to spatial decision analysis, a more
vivid graph expression than DSS, and thus enable the user to
assess the implications of the trade-offs between alternatives
clearly. Its primary functions are to (a) provide the mechanisms
for interactive input and manipulation of large volumes of
spatial data; (b) allow representation of the complex spatial
relationships and structures that are common in spatial data,
including analytical techniques that are unique to both spatial
analysis and modeling; (c) provide output in a variety of spatial
forms; and (d) facilitate decision-making and improve the
effectiveness of the decision made (Turban, 1988). SDSS are
explicitly designed to provide the user with an interactive
decision-making environment that enables geographic data
analysis and spatial modeling to be performed in an efficient
and flexible manner. Basic component of SDSS include control
187
unit, database (management), model base (management),
knowledgebase (management) and user interface, with spatial
database different from common database in DSS (Fig. 1).
Until now many a SDSS have been developed and applied in
different areas by various researchers. Arentze and
Timmermans (2000) described the architecture of a spatial
decision support system and an illustrative application, to
generate retail plan and assess its impact. Mchiael (2001)
presented a GIS-based decision support system prototype
intended for use by public housing authority administrators and
planners designing policy for housing mobility programs. Vacik
and Lexer (2001) researched the development and application
of a spatial decision support system (SDSS) for silvicultural
planning in forests managed for sustained yield of water
resources. Keenan (1998) developed SDSS to server vehicle
routing. Zergera and Smith (2003) emphasis the importance of
knowledge in using GIS for real-time disaster decision support.
In all those cases, the component of database and model base is
a prerequisite, while knowledgebase is optional since
knowledge places the base for intelligent decision-making that
may not required.
o
Data Model Knowledge |
base : base [E base d
A |
^ Sub-system of ^
‘Sub-system of” Sub-system of ^v
V database modelbase knowledgebase |
. management Ls mem management
l l y
Control unit |
J
/
Use interface
Fig. 1 Basic structure of SDSS
2.2 GIS and GIS Use in Agriculture
The GIS technology has come a long way in the past decade
and continues to evolve, with the basic function as spatial data
management. New application areas have been found, including
agriculture, forestry, hydrology, resource management, and
coastal resource management. Those areas benefit a lot from the
development of GIS. In addition, new products have appeared
in the marketplace. What more, dramatic improvements
continue in the capability of hardware and software operating
platforms; and large volumes of data sets have become
available. GIS technology has grown rapidly to become a
valuable tool in the analysis and management of spatial
ecological problems.
It is not new for GIS to be used in Agriculture. Since the
Canada Geographic Information System or CGIS, generally
acknowledged as the first GIS system (Peuquet, 1977), GIS has
been applied in resource planners and decision-makers with a
set of tools to analyze spatial data effectively. Agricultural
resource plan, agricultural land assessment, etc.. are also among
the areas that GIS can provides. These areas can be classified as
macro application since large area is usually covered. A more
