A strategy for the automatic synthesis of plant operating procedures, through the integration of artificial intelligence and mathematical modeling tools is discussed. This "predicative/numerical" modeling paradigm integrates qualitative and quantitative aspects of the problem in a same framework. It uses a "process-centered" qualitative modeling approach together with MILP optimization techniques to automatically generate operating strategies. The operation procedures network (OP-Net), formulated in terms of processes from a generic library, describes abstract (hardware-independent) processing paths leading to the desired operational goals. In this approach, the model assumptions, conservation expressions and data are described at the same level of abstraction, using the same formal representation. This permits the direct, automatic generation and modification of operation procedures in situations involving changes in basic assumptions and initial scenarios. Thus, flexibility, an essential feature of batch processing, is safeguarded.