This paper focuses on job shop scheduling problems in a cooperative environment. Unlike classical deterministic approaches, we assume that jobs are not known in advance but occur randomly during the production process, as orders appear. Therefore, the production schedule is adapted in a reactive manner all along the production process. These schedule adaptations are made according to a cooperative approach, that is the major originality of this paper. Each resource manages its own local schedule and the global schedule is obtained by point-to-point negotiations between the various machines. We also suppose that local schedules are flexible since several alternative job sequences are allowed on each machine. This flexibility is the key feature that allows each resource, on the one hand, to negotiate with the others and, on the other hand, to react to unexpected events. The cooperative approach aims at ensuring the coherence between the local schedules while keeping a given level of flexibility on each resource.