Multi-objective/Multi-criteria Analysis Module
For the task of analyzing the conflicting goals/objectives of the
different decision levels and disciplines involved in watershed
management.
Inference Engine
Manipulates the knowledge represented in the knowledge base
to infer a solution to the problem(s) described by the
information in the databases.
Rule-based system
This is the knowledge representation scheme. It implements the
concept of data-driven and goal-directed reasoning. The rule-
based system contains the knowledge about a particular domain
that has been acquired from domain experts.
DSS Database
This database contains specific data (initial and derived data
when solving the problem) which describe all facts that are
known about the problem. Data may be the result of: some GIS
operations; data automatically collected by a data acquisition
system; sets of answers provided by the end user in response to
Application Databases (DSS DB) Level 4
> à 4
Federated 0.0. Scale Level 3
Schemas levels
O.O. schema
Sharable data in
O.O. schema
| Sharable data in
Sharable data in
[^ 0. eg Level 2
1 Shareable data Shareable data Shareable data Level 1
i (Schema 1) (Schema 2) (Schema 3) ove
Schemas
Elementary Elementary Elementary Level 0
DBI DB2 DB3
Databases”
Figure 4 Multi-level federate database
the queries posed by the inference engine; or data exported from
remote databases.
Object schema
The object schema describes the entities in the domain (in terms
of classes, objects, properties, methods, inheritance, etc.). It also
enables the retrieval of the object properties from the SDSS
database using the Object-Oriented data model. The object
schema provides an integration layer that supports the
requirements of the different SDSS components mentioned
above.
GUI
It provides users with an intuitive interface for interacting with
the SDSS and its various components. The basic objective of the
GUI is to allow users to access GIS, RS, application programs,
rule base and multi-objective analysis components from a single
interface.
In this section, the components of the MLDSS were shown.
First the overall architecture of the MSLDSS were explained,
then the components of the SDSS were presented. The
supporting databases of such an architecture are heterogeneous
in many respects. The next section shows how the data models
are organized and structured in the MLDSS.
4. HETEROGENEITY IN DISTRIBUTED GISS
[Worboys, M., et al., 1991] classified semantic heterogeneity as
generic and contextual. The earlier occurs when different GIS
applications use different data models of representing their
spatial information. For example one may use a layer-based
approach while a another may use an object-based approach,
The contextual heterogeneity occurs when the semantics of
schemes depend upon the local conditions at particular GIS. For
example two spatial databases contain two different objects
which have two different meanings, though they refer to the
same real world entity, e.g., agricultural fields in environmental
database are different from those in a cadastral database.
[Spaccapietra, S., et al., 1991] listed 4 classes of heterogeneity
or conflicts: semantic conflict, descriptive conflict, data model
conflict, and structural conflict. The semantic conflict occurs in
the situation where two sets of objects from two schemes are
representing sets of real world entities which are related by a set
comparison operators other than equality. Descriptive conflict
occurs when two database objects, representing the same real
world entity, are described with different sets of properties. Data
model conflict is the situation where two schemes are defined
with two different data models, e.g., relational and object
oriented models. The situation where two related objects are
represented using different data structures is called structural
conflict. For example a designer represents a component X of an
object O either by creating a new object type X or add it as a
property of O.
A relatively similar classification of types of database
heterogeneity is presented by [Saltor et al, 1993]. They
provided more comprehensive classification of heterogeneity to
which we are more inclined. Their classification has three
aspects: syntactic, schematic, and semantic. Descriptive and
structural conflicts are equivalent to schematic heterogeneity,
while the data model conflict is equivalent to syntactic conflict.
1. Syntactic: each database may be implemented in a different
DBMS with a different data model, e.g., relational model Vs
object oriented model. Syntactic heterogeneity is also related
to the geometric representation of geographic objects, e.g.,
raster and vector representations.
2. Schematic: where objects in one database are considered as
properties or metadata in the other, or object classes of the
same real world entity have different hierarchies and
descriptors in different databases.
3. Semantic: a real world entity may have two different
meanings in their underlying databases in order to serve
various applications, giving as a consequence semantic
conflicts. For example a road network in a GIS for
transportation has different semantics from that in a GIS for
668
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
