Show simple item record

Genetic Algorithms for timetable Generation

dc.contributor.authorWalusungu, Gonamulonga Gondwe
dc.date.accessioned2017-07-20T09:19:35Z
dc.date.available2017-07-20T09:19:35Z
dc.date.issued2014-12-15
dc.identifier.urihttp://repository.aust.edu.ng:8080/xmlui/handle/123456789/590
dc.description.abstractTimetabling presents an NP-hard combinatorial optimization problem which requires an efficient search algorithm. This research aims at designing a genetic algorithm for timetabling real-world school resources to fulfil a given set of constraints and preferences. It further aims at proposing a parallel algorithm that is envisaged to speed up convergence to an optimal solution, given its existence. The timetable problem is modeled as a constraint satisfaction problem (CSP) and a theoretical framework is proposed, which guides the approach used to formulate the algorithm. The constraints are expressed mathematically and a conventional algorithm is designed that evaluates solution fitness based on these constraints. Test results based on a subset of real-world, working data indicate that convergence on a feasible (and optimal/Pareto) solution is possible within the search space presented by the given resources and constraints. The algorithm also degrades gracefully to a workable timetable if an optimal one is not located. Further, a SIMD-based parallel algorithm is proposed that has the potential to speed up convergence on multi-processor or distributed platforms.en_US
dc.description.sponsorshipAUSTen_US
dc.language.isoenen_US
dc.publisherAUSTen_US
dc.subjectGenetic Algorithmsen_US
dc.subjectWalusungu Gonamulonga Gondween_US
dc.subjectProf Lehel Csatoen_US
dc.subject2014 Computer Science Thesesen_US
dc.titleGenetic Algorithms for timetable Generationen_US
dc.typeThesisen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

  • Computer Science105

    This collection contains Computer Science Student's Theses from 2009-2022

Show simple item record