Tech Reports

Tech Report kmi-02-03 Abstract

The Task Ontology Component of the Scheduling Library
Techreport ID: kmi-02-03
Date: 2002
Author(s): Dnyanesh Gajanan Rajpathak

Scheduling is a ubiquitous task spanning over many activities in day to day life. Usually, a scheduling problem comes in a variety of flavours, which makes it a hard problem both in theory as well as in practice. The scheduling task concerns with the assignment of jobs to the resources and time ranges within a pre-defined time framework while maintaining various constraints and satisfying requirements. Due to the diverse nature of scheduling problem the nature of its main building blocks differs according to the target domain. Such a changing nature of target domain increases the overall time and cost required for building an application system. An ontology can be seen as a reference model that describe the various entities that exist in universe of discourse along with their properties. These entities can be individuals, classes, relationships, and functions. In sum anything that can be useful for describing the classes of task in hand. In this report we propose generic task ontology for constructing scheduling applications. The proposed task ontology is generic to emphasise that it is both domain as well as application independent. We refer to it as task ontology such that it describes the classes of scheduling task independent of the various ways by which this task can be solved. We envisage moving beyond the current brittle approaches to the system development by providing the firm theoretical as well engineering foundations for the various classes of knowledge-intensive scheduling applications.