DocumentCode
1624789
Title
Novel hybrid PSO algorithms with search optimization strategies for a University Course Timetabling Problem
Author
Oswald, Cyril ; Deva Durai C, Anand
Author_Institution
Indian Inst. of Inf. Technol., Design & Manuf., Chennai, India
fYear
2013
Firstpage
77
Lastpage
85
Abstract
A multi-dimensional assignment problem in which Students, Faculty and Subjects together as events are scheduled to suitable timeslots and available classrooms, is termed to be University Course Timetabling Problem (UCTP). Previous literature studies reveal that getting the best fitness function value with generating an optimized timetable is not considered to a great extent. This paper proposes a novel method of solving the UCTP through various Hybrid Search Optimization algorithms combined with Particle Swarm Optimization (PSO). This approach is implemented for the real time data taken from the Faculty of Computer Science and Engineering of a typical Indian University. Our experimentation clearly confirms by giving promising results that, getting a near or sub optimal solution is highly feasible through various hybrid algorithms.
Keywords
educational courses; educational institutions; particle swarm optimisation; search problems; Faculty of Computer Science and Engineering; Indian University; UCTP; classrooms; fitness function value; hybrid PSO algorithms; hybrid search optimization algorithm; multidimensional assignment problem; near-optimal solution; optimized timetable generation; particle swarm optimization; search optimization strategies; sub-optimal solution; timeslots; university course timetabling problem; Annealing; Equations; Mathematical model; Fitness Function; Hybrid Particle Swarm Optimization; Particle Swarm Optimization; University Course Timetable Problem;
fLanguage
English
Publisher
ieee
Conference_Titel
Advanced Computing (ICoAC), 2013 Fifth International Conference on
Conference_Location
Chennai
Print_ISBN
978-1-4799-3447-8
Type
conf
DOI
10.1109/ICoAC.2013.6921931
Filename
6921931
Link To Document