Back to publications overview.
Keywords: Manufacturing, Job Shop, Project Planning, Intelligent Agents, Discrete Simulation.
A simulation approach is presented for planning and scheduling a flow of complex jobs for job shop like production systems. Machines may have their own specific restrictions and properties such as relative production speed, set up characteristics and scheduling rules. A production job consists of a set of tasks represented by a directed activity network in which an activity is defined as a single task to be processed on a machine. As a consequence a machine may also be an assembly station. The task duration may be stochastic having any probability distribution. The task flow and the task selection for scheduling is governed by agents: Each machine group acts as an agent combining tasks and machines. Each task acts as an agent guarding the proper network sequences of its job by determining when it is ready for releasing and, after that, controlling its own scheduling priority. A job acts as an agent repeatedly updating critical path analysis for its tasks. For that purpose a separate critical path simulation model is used. Crucial for the modelling the "process approach" used. The model is generic with respect to job shop configuration e.g. number of machine groups and number of machines per group and also with respect to the job configuration.