Abstract
Creating a timetable is a complex task and an interesting problem for testing differentheuristic optimization methods. In this master thesis, based on real-world data from practice,
a solution representation including fitness evaluation is developed and used for comparing
optimization methods.
There is already a lot of literature on this topic, and even competitions exist, so many
approaches are available. The solution developed in this master thesis is tailored specifically
to a school and its unique constraints and challenges. This different perspective—not aiming
to find a highly generic solution—gives this thesis the opportunity to explore whether it is
necessary to apply complex heuristic methods for such highly customized solutions.
In this thesis, for example, Random Search is compared with a Genetic Algorithm, which
might feel like comparing a trowel to a bucket-wheel excavator. In addition to these two
methods, Simulated Annealing, Variable Neighborhood Search, Iterated Local Search,
Greedy Randomized Adaptive Search, Tabu Search and Local Beam Search are also tested
for their suitability in optimizing and minimizing errors.
Date of Award | 2024 |
---|---|
Original language | German (Austria) |
Supervisor | Erik Pitzer (Supervisor) |