University course timetabling problem is a complicated problem and finding a computer-aided solution for it was a subject to work for many years. To solve this problem, we must assign courses to timeslots with respect to hard and soft constraints. Hard constraints are those which must be necessarily met (some of them could be neglected with high costs). Our aim is to meet as many soft constraints as possible. In this paper we present local search algorithms to improve a table which meets hard constraints, and is provided by a searching algorithm for CSPs. Our results show that the proposed method is very suitable for real data with a large size and complex constraints.
)
