Intelligent scheduling framework

Abstract

This dissertation describes the main properties of scheduling and planning problems. A standard representation of the problems was designed and appropriate data structures and data stores were designed and implemented. An application programming interface (API) was implemented for the system user to make use of while implementing dedicated scheduling algorithms. A framework was then designed around this generic problem description and application programming interface. Several examples were then implemented to show the concept and flexibility of the system. The problems were chosen from totally independent domains, to show the generic nature of the framework. In this part of the dissertation, algorithms from different areas of artificial intelligence were implemented and tested. This was also done in order to exploit the generality of the framework, in order to solve problems from different domains, using techniques from different study areas. The most important issues of this dissertation are the following: accessing data in an abstract and optimal way, generic implementation of all system components with reusability in mind, transforming problems from different domains to the generic representation of the framework, and using the system in a distributed environment such as the web.