The thesis aims to solve the high school timetabling problem, where lessons must be assigned to time slots for specific classes and teachers. Each lesson has an assigned teacher who can teach at most one subject at a time, and lessons within the same class cannot overlap. Constraints include consecutive lesson assignments (with a mandatory lunch gap), division of students into groups, and bi-weekly (odd/even) cycles. These hard constraints are accompanied by soft constraints on the exclusivity of lessons within a day, the similarity between odd/even weeks, and the avoidance of undesirable early or late assignments.

The thesis will propose and implement a solution motivated by a real-life problem. The approach will follow the large neighborhood search framework, specifically adapting the slack induction by string removals algorithm, originally proposed for vehicle routing, to the timetabling domain. The implementation will be evaluated using real-life data from the High School of ICT, Postal Services, and Finance in Brno.