Agricultural University of Athens
Σχεδιασμός και ανάπτυξη συστημάτων αποφάσεων βασισμένων σε ολοκληρωμένα έμπειρα γεωγραφικά συστήματα βάσεων δεδομένων
Abstract
dc:descriptionThe main goal of this dissertation is the description of the design and implementation of an Integrated Geographical Expert System (IGES) on a Database which integrates under a common environment the Expert Systems with the Geographical Information Systems, taking advantage of the capabilities of Databases.The proposed ΙGES, is implemented by using the following structural elements: a. the Database Management System (DBMS), b. the Expert System (ES), c. the Geographical Information System (GIS) and d. the appropriate Graphical User Interface (GUI).Current research’s effort basic motive is to present the possibility of independent and ostensi-bly heterogeneous environments (the mentioned before structural elements) to interoperate, by utilising data and knowledge, with final objective the projection of the results with the most suitable way.Aim of the developed IGES is the information provision and solution proposal in all the activi-ties of basic agricultural applications to the involved in the respective subjects researchers, consultants, producers (professionals and amateurs) as well as to all end users. IGES has been fully applied in a significant for Greece agricultural sector, as is the apiculture. The un-abridged implementation covers all the apiarian activities. Concretely, the system helps in the determination of optimal conditions (space, time, remaining parameters) for the beehive shift-ing aiming at the qualitative and highest honey production.In knowledge area, we describe the architecture, representation and operation of a Hybrid Expert System (HybES) in the sense of incorporating methodologies of Expert Systems, relational databases, fuzzy logic data sets and criteria. More specifically, the introduced HybES is a Rule Based Expert System, deployed on a relational database with its rules based on the Object-Attribute-Value (O-A-V) triplet representation model. Every rule is valid for a specific time period and, consequently, the whole system is valid at that certain time periods, using different valid-time versions of knowledge. The values of the rules (O-A-V triplets) are related to fuzzy sets, handling linguistic variables. Each one of these linguistic variables has the ability to depend on one or more criteria affecting it. The system infers using either the forward chaining (applying ‘rules forward’) or the backward chaining (a bottom-up method, applying the rules in the opposite way). Using a simple technique, the working memory of the system is split into two parts (Conditions’ and Rules’ Working Memory), which are used according to system’s needs.The developed IGES uses an embedded ES for the utilization of knowledge, while the devel-opment using HybES constitutes a future work aiming at the obtainment of independence, but contemporaneously the interoperability of miscellaneous structural parts.We notice that the most part of this dissertation has been published in the scientific journals and proceedings referred in page 28.
Degree
thesis:*- Grantor dc:publisher
- Agricultural University of Athens
- Year dc:date
- 2006
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Filis, Ioannis
- Φίλης, Ιωάννης
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/21669