SUNConferences, COMA '13

Font Size: 
Production Flow Control using Biased Minimum Feedback
Anthony John Walker, Glen Bright

Last modified: 2013-08-21


Manufacturing flow control in high-variety production attempts to adjust execution sequencing so as to synchronise throughput rates over reconfigurable manufacturing resources. The flow control problem is especially concerning when the time to setup a resource in the required configuration for imminent production requests is significant. This results in a complex scheduling and decision making problem. The paper presents an event driven control formalism in which the optimal batch sizing problem is collapsed onto a non-deterministic finite state automata driven by Biased Minimum Feedback (BMF). It is shown that optimal batching occurs as a result of stochastic event trajectories in a finite state Markov process. Properties and characteristics of the metrics used in BMF implementation are studied over varying time horizons. The stability of BMF is proven using an equivalent discrete time system and the notions of positive definite matrices. A simulation study on the control of a reconfigurable manufacturing cell uncovers work-in-process levels which optimise setup frequency and assembly synchronisation under high-variety production.

Full Text: PDF