On the Boundary of Planning and Scheduling: A Study

Planning and scheduling are areas attracting an unceasing attention of computer science community. However, despite of their similar character, planning and scheduling problems are usually handled independently using different methods and technologies. While Artificial Intelligence technology is more relevant for planning problems, Operations Research has a long tradition in studying scheduling problems. Recently, Constraint Programming brings a fresh breeze to both areas as it allows a combination of various methods to solve the planning and scheduling problems within single declarative framework. In the paper we give a study of mixed planning and scheduling approach to solve the scheduling problems in complex production environments. We analyse three conceptual models and we compare their advantages and limitations from the mixed planning and scheduling point of view. We also give an industrial background to justify the features required in the models. The models were studied within the VisOpt project whose goal is to developed a generic scheduling engine applicable to various complex production environments. However the results can be applied to other (non-production) problem areas where mixed scheduling and planning capabilities are desirable.