An investigation of evolutionary algorithms for automated student project allocation
Otomatikleştirilmiş Öğrenci Proje Dağıtımı için Evrimsel Algoritmaların Araştırılması
- Tez No: 788168
- Danışmanlar: DR. LEANDRO L. MİNKU, DOÇ. DR. NİR PİTERMAN
- Tez Türü: Yüksek Lisans
- Konular: Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Bilim ve Teknoloji, Computer Engineering and Computer Science and Control, Science and Technology
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2017
- Dil: İngilizce
- Üniversite: University of Leicester
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Enformatik Ana Bilim Dalı
- Bilim Dalı: Bilgisayar Bilimleri Bilim Dalı
- Sayfa Sayısı: 62
Özet
Özet yok.
Özet (Çeviri)
The Student Project Allocation Problem (SPAP) is a NP-hard optimisation problem that can be thought of as a subset of resource allocation problems. There are various versions of SPAP that attempt to assign students to projects or supervisors to projects and alike. In this study, the problem consists of allocating projects to students, assigning a supervisor for each allocation, and finally appointing a second marker for each student-project-supervisor matching in such a way that it satisfies predefined constraints and rules. The genetic algorithm method is preferred to solve the problem because the problem is available to be encoded in genetic representation and has straightforward evaluation criteria. This study aims to investigate which genetic algorithm operators are best for improving the fitness of the solution and achieving consistency at finding feasible solutions. Different strategies of different operators have been implemented and tested on the problem to observe their performances. During the experiments, four different datasets representing different scenarios have been used. These datasets are derived from two real-life datasets obtained from the Department of Informatics, University of Leicester. In this thesis, the ways in which the problem can be represented, the operators can be implemented, and the problem can be converted into an optimisation problem are explained. To facilitate an environment for experiments, a tool has been developed which is able to be utilized for solving real-life allocation problems also. The outcomes of the experiments have been interpreted by using statistical hypothesis tests and the findings of these tests have been represented. The outcomes of the experiments illustrate that truncation, ranking and tournament parent selections are better comparing the rest and the uniform crossover is the best by far. As for the mutation mechanism, applying mutation over both project and supervisor provides better allocations.
Benzer Tezler
- A knowledge discovery approach to urban analysis the Beyoğlu preservation area as a data mine
Kent analizinde bir bilgi keşfi yaklaşımı bir veri madeni olarak Beyoğlu kentsel koruma alanı
AHU SÖKMENOĞLU SOHTORİK
Doktora
İngilizce
2016
Mimarlıkİstanbul Teknik ÜniversitesiMimarlık Ana Bilim Dalı
PROF. DR. GÜLEN ÇAĞDAŞ
PROF. DR. İKBAL SEVİL SARIYILDIZ
- Evolutionary design assistants for architecture
Mimarlık için evrimsel tasarım asistanları
N. ONUR SÖNMEZ
Doktora
İngilizce
2015
Mimarlıkİstanbul Teknik ÜniversitesiMimarlık Ana Bilim Dalı
PROF. DR. ARZU ERDEM
PROF. DR. İKBAL SEVİL SARIYILDIZ
- Gezgin satıcı probleminin çözümünde genetik algoritmanın parametrelerinin incelenmesi
Investigation of parameters of genetic algorithm in the solution of traveling salesman problem
MERYEM PULAT
Yüksek Lisans
Türkçe
2017
EkonometriDokuz Eylül ÜniversitesiEkonometri Ana Bilim Dalı
PROF. DR. İPEK DEVECİ KOCAKOÇ
- Energy aware endurance framework for mission critical aerial networks
Güdümlü havasal ağlar için enerji farkında endürans modeli
YUSUF ÖZÇEVİK
Doktora
İngilizce
2019
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
DOÇ. DR. BERK CANBERK
- An Investigation of Evolving Deep Time Variant Network Controllers on Complex Fitness Landscapes
Başlık çevirisi yok
KAAN AKINCI