The presented work deals with the design of production systems in uncertain context. The design of such systems can be interpreted as an optimization problem that consists to find a configuration optimizing certain objectives and respecting technological and economical constraints. The production systems studied in this thesis are the assembly and transfer lines. The first one is the line that can be represented as a flow-oriented chain of workstations where, at each workstation, the tasks are executed in a sequential manner. The second is a particular line that is composed of transfer machines including several multi-spindle heads where the tasks are executed simultaneously. At first, we describe different approaches that permit to model the uncertainty of data in optimization. A particular attention is attracted to two following approaches: robust approach and sensitivity analysis. Then, we present three applications: the design of assembly and transfer lines under variations of task processing times and the design of an assembly line with interval task processing times. For each application, we identify the expected performances as well as the complexity of taking into account the uncertainty. Thereafter, we propose some new optimization criteria in adequacy with the introduced problematic. Finally, resolution methods are developed to solve different problems engendered by these criteria.
