Exact and heuristic methods for solving the Car Sequencing Problem

In this thesis, I present several different methods for solving the Car Sequencing Problem. The process of scheduling vehicles along a production line has to take several constraints into account which are defined by the body shop, the paint shop and the assembly shop. I attached importance especially to the paint shop constraints, because no violations are allowed here. Beside the enhancement of an Integer Linear Program (ILP), I developed an alternative formulation. I give a review about the Variable Neighbourhood Search (VNS) and combine this metaheuristic with the exact methods to achieve good heuristic results in relatively short time. In addition, a newly developed heuristic using exact methods is presented which is also used as starting heuristic for VNS. I compare the results obtained by my algorithms with results published in the literature using two sets of benchmarks. One set is taken from the CSPlib, an publicly available set of problems. ROADEF and the car manufacturer Renault published the other set of instances for the ROADEF Challenge 2005. In contrast to the CSPlib instances, the ROADEF instances include constraints defined by the paint shop. In some cases, I was able to half the computation time for the CSPlib instances using exact methods. For the instances defined by ROADEF some new best solutions were obtained.