Dynamic hierarchical scheduling strategies for manufacturing systems

The purpose of this research is to develop an effective scheduling strategy to dynamically schedule the production floor of a manufacturing system in which jobs arrive continuously over time. To take advantage of the added machine routing flexibility and the existence of natural levels of decision makers, the scheduling function is divided into two decision levels consisting of job-to-machine assignment strategy and the job sequencing strategy.; A two level scheduling strategy has been developed in which the job-to-machine assignment decisions are made by the cell controller and the actual start and completion times of the jobs are made by the machine supervisors. Two methods of implementing the two level scheduling strategy have been proposed; a two level heuristic scheduling strategy and an iterative schedule generation using performance forecasts. The effectiveness of the multilevel dynamic scheduling strategy for jobs that arrive continuously with prescribed due dates is studied.; The first implementation of the two level scheduling strategy is using a two level scheduling technique in which heuristic rules are used to determine the job-to-machine routing and job sequencing is proposed. The goal is to gain experimental experience of the use of a two level scheduling strategy. The allocation of an individual job to an individual machine is made using a "fair" job-to-machine assignment strategy called Minimum Job In Queue (MINJQ). The routing strategy is combined with one of the four representative dispatching rules chosen in specifying the job sequence. These dispatching rules are: First Come First Served (FCFS), Shortest Processing Time (SPT), Earliest Operation Due Time (EODT), and Earliest Final Due Date (EFDD).; The second implementation of the two level scheduling strategy is the application of the leader-follower strategy from the theory of control systems to the decision making processes in manufacturing systems containing several hierarchical decision levels. The generation of strategies for hierarchical scheduling of manufacturing systems using performance forecasts from locally controlled subsystems is described. The performance forecasts are used to iteratively measure the reactions of the subsystems to the tentative coordination strategy provided by the leader to arrive at a collaborative final system decision. A dynamic scheduling technique in which performance forecasts are used as feedback for the cell controller is presented. The scheduling strategy is capable of handling dynamic job arrivals and the need for job rescheduling due to machine failures and material shortages.; The effectiveness of the two level heuristic scheduling strategy and the iterative scheduling strategies have been studied using several simulations. The sensitivity of the scheduling methods due to job due date variations is also presented.