Research On Ontology-based Timetabling Problem

TTP (Time Table Problem) belongs to a research topic widely applied. This paper deals with class schedule, which is the typical application of such issue.With the increasing number of students, class schedule has been cumbersome and important teaching work. Therefore how to schedule class rationally is crucial for education management of each school. Class schedule is a constrained, multiple objective combinational optimization problem. And it proves an NP-complete problem. Various solutions have been proposed to handle this problem, however for the complexity and variety of class schedule, perfect solution has not been obtained for years.This paper presents a novel scheme to alleviate the issues mentioned above by using the ontology method. Firstly, class schedule problem is treated as resource allocation problem. That is the related resources such as teachers, students and classes should be dispatched as school requirement. After the model of this problem is abstracted, the constructive steps are followed as existing regulations to establish class schedule ontology. Secondly, based on description logic, the involved resource knowledge and relationships among them are described in OWL DL. Semantic Web Rule Language (SWRL) defines the solution to collision, and the constraints are also expressed in SWRL. In this way, the preliminaries including knowledge domain and regulation domain are made for reasoning. Thirdly, ahead of reasoning, the class schedule problem ontology is re-categorized and collision checked by Racer reasoning machine to make sure that there is no contradiction between facts. Lastly, Jess reasoning machine is employed to connect field ontology and reasoning engine. The instance of ontology is inverted into the fact in corollary engine, based on this, reasoning is carried out.Tool Protege is adopted as tool for class scheduling, and the demonstration of the above process is performed. And the feasibility of our approach to solve multiple objective constraints class schedule problem is proved. Thus knowledge share and reusage is implemented. Class schedule system obtained by our proposed method can be better maintained and applied.