Control of push and pull manufacturing systems

In this thesis, two common models for manufacturing systems are studied: the push model and the pull model.In the push model, raw parts are input by a push mechanism into the system. For such models, the stability and performance of several classes of scheduling policies are analyzed. Each unit of a given part-type requires a predetermined processing time at each of several machines, in a given order. A set-up time is required whenever a machine switches from processing one part-type to another.For nonacyclic push systems, clearing policies need not be stable. By modifying Clear-A-Fraction (CAF) policies slightly, a policy which stabilizes all systems is obtained.For a single machine push system with constant demand rates, it is shown that, for any Generalized Round Robin (GRR) scheduling policy, the buffer level trajectories of each part-type converge to a periodic steady state trajectory. The periodic trajectory is computed by solving a set of simultaneous linear equations. The performance of Round Robin (RR) policies, a subclass of GRR policies, is compared with a lower bound on achievable performance.In the second part of this thesis, a pull model of manufacturing systems is analyzed. First a multimachine flow shop producing a single part-type, that attempts to satisfy exogenous constant demand rates, subject to nondecreasing buffer costs, is considered. The objective is to determine the optimal control for the production rate at each of the machines in the system, in order to minimize the total buffer holding cost incurred while emptying the system.When there is an initial surplus of the finished product, the optimal control is trivially determined. In this case, the result generalizes to re-entrant manufacturing systems. If there is initially a shortage of the product, then it is shown that the optimal control can be characterized by N "deferral" times, where N is no more than the number of machines. The optimal control problem is determined by solving a set of N quadratic programming problems.Finally, a re-entrant line producing a single part-type is analyzed. The optimal control is determined, for the case in which the last machine is the bottleneck.U of I OnlyETDs are only available to UIUC Users without author permissio