In this thesis, we deal with modeling and solving various problems including vehicle routing and scheduling. We propose approximate methods that aim to solve problems qui- ckly and e ciently. Five problems are addressed. The rst one is the Resource-Constrained Project Scheduling Problem (RCPSP) for which a multi-commodity ow approach is intro- duced. We also consider methods dealing with extensions of this problem ( nancial or time constraints). The second one is a two dimensional Orthogonal Packing Problem (2OPP) with the resolution being based on its RCPSP relaxation. The third one is the Stacker Crane Problem (SCP). It is a pickup and delivery problem: goods must be transported from starting points to various destinations with a eet of vehicles. In the SCP, a single one-capacity vehicle is available. We propose original tree based models and algorithms for the preemptive case. The fourth one is a Dial-a-Ride Problem (DARP) with nancial constraints. We solve this problem through an insertion heuristic using a constraint pro- pagation technique. The fth one combines routing problem and two dimensional packing. It is the 2L-CVRP in which items have to be delivered to customers. A GRASP ELS framework is introduced for this problem. Experimental results show the e ectiveness of the proposed methods.
