Abstract (EN)

In this paper, we consider the problem of evaluating the worst case performance of flexible solutions in single machine scheduling. A flexible solution is a set of schedules following a structure defined by a partial order of jobs and a type of schedules. here, we consider active and non-delay schedules. The proposed structure allows the decision maker to dispose of some flexibility (a set of schedules instead of one schedule) that can be used on-line to absorb the impact of data disturbances related to, for example, job arrival, tool availability or machine breakdowns. The performance is a function of (i) the best case performance providing the decision maker a lower bound of the performance, (ii) the worst case performance giving him a guarantee about how poorly the solution may perform, in both cases when following the considered structure. The worst cases correspond to new scheduling problems, in which a regular criterion must be maximized. We present new algorithms and/or new computational complexity results, in presence of: one machine, different release dates, precedence constraints and various criteria to be maximized. || Nous nous intéressons dans ce papier à l'évaluation de performance dans le pire des cas pour des solutions d'ordonnancement flexibles pour des problèmes d'ordonnancement à une machine.