The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
1081
Figure 1. Illustration of Service Oriented Architecture (SOA)
3. DECISION SUPPORT SYSTEM FOR DISASTER
MANAGEMENT
Disaster management is a very complex process. It contains not
only sensing, monitoring and detection of pre-disaster, but also
response, rescue and recovery of post-disaster. That is to say,
disaster management covers all four stages including
preparedness, mitigation, response and recovery (Chen et al.,
2006; Fiedrich and Burghardt, 2007).
As stated above, disaster management involves a variety of
disaster response entities, such as ambulance team, fire brigade,
doctors and police office, etc. On the one hand, each entity has
its own duty and responsibility. On the other hand, all the
entities involved must cooperate and collaborate closely. It
means that their interactions must be considered carefully.
Multi-agent system and rule-based expert system play critical
roles in the implementation of decision support for disaster
management. More details will be discussed in the following
section.
3.1 Multi-Agent System (MAS)
Multi-agent system provides a mechanism and platform to
support the cooperation and collaboration among agents. Within
a multi-agent system, the task is split into individual subtask.
Each agent will take charge of one specific subtask. Finally, the
task can be achieved through the cooperation and collaboration
among agents. Thus the very nature of multi-agent system does
fit the requirements of disaster management. In the context of
disaster management, each disaster response entity can be
regarded as an agent, and then each agent can be assigned to a
subtask. Therefore, the process of disaster decision making can
be transformed into the cooperation and collaboration of
corresponding agents. The representation of disaster response
entities through multi-agent system is shown in Figure .
Figure 2. Representation of Relevant Entities using MAS
3.2 Rule-Based Expert System
Rule-based expert system provides a nature way to capture and
represent the expertise and professional knowledge related with
disaster management and emergency response. Within a rule-
based expert system, domain knowledge will be represented as
a set of rules. A rule presents specific description of how to
solve a given problem. Each rule includes the IF...THEN
structure, i.e., IF <condition> THEN <action>. When the
condition of IF part comes into true, the rule is said to fire and
then the action of THEN part will be executed. The basic
workflow of a rule-based expert system is shown in Figure 3.
Figure 3. Basic Workflow of Rule-Based Expert System
When using rule-based expert system, a fundamental question is:
where does the domain knowledge come from and how to
capture them into knowledge base?
For disaster management and emergency response, the source
of relevant knowledge includes human expertise and
documented data and information. The knowledge obtained
from human experts and documented data and information can
be captured into knowledge base through knowledge
engineering. Figure 3 shows the process of creation of disaster
management knowledge base.