Use Historic Information as Self Organizing Resources for an Intelligent Timetable Scheduling System

Abstract

An application of office automation is proposed in this paper. We use past historic timetable data as templates to intelligently adjust a better evolution of a set of more usable timetables. We develop a two staged system: First of all, a feasible set of timetables were generated via blank-filled technique. These timetables merely satisfied users’ basic constraints. Then, historic tables and teachers’ preferences were feed to the system in order to fine tune these tables. The intelligent system has the capability to self-organize these set of timetables. We conduct a questionnaire for volunteers including students and instructors. Two different timetable sets (before/after the tuning process) were given to these testers. Fixed sized sample of candidates were randomly picked regularly. And the chosen testers have to grade the tables depending on their satisfactoriness toward the timetable they received and presumably they are going to use that timetable later on. A two sample Welch t-test on the average of testers’ score was given. Under 90% of confidence level, there is a significant difference between the satisfactory mean of these tables. Auto-tuned timetable bears comparison with the manual work of the current assistant.