Abstract (eng)
This paper creates an integer programming approach to the real-world problem, which partly bases on course timetabling problem, as well as includes unique formulation required by the case. Precisely, the model presented in this paper aims to maximize the amount of monthly held lectures that are assigned according to students’ preferences. A presented timetable also depends on a number of hard (e.g. no more than one student in the same plane at the same time) and soft constraints (e.g. no assignment of expensive plane type for the beginner courses) in order to consider diverse restrictions set by a school and hence to design a realistic timetable.
The problem was tested on three different data sets with a help of mixed integer program solver (FICO Xpress). The formulation developed in this thesis has shown good practical results, especially for smaller problems. Medium and large instances were subject to simplification of the formulation due to the need of improvement in computational runtime and solution. Solutions to large problems require further investigation of possible relaxations, as well as the use of non-exact methods in order to improve runtime and method optimality.