Opportunities for Improved Statistical Process Control

Abstract

Our Bayesian dynamic programming model builds on existing models to account for inspection delay, choice of keeping production going during inspection and/or restoration, and lot sizing. We focus on describing how dynamic statistical process control (DSPC) rules can improve on traditional, static ones. We explore numerical examples and identify nine opportunities for improvement. Some of these ideas are well known and strongly supported in the literature. Other ideas may be less well understood. Our list includes the following: Cancel some of the inspections called for by an (economically) optimal static rule when starting in control (such as at the beginning of a production run and following a restoration). Inspect more frequently than called for by an optimal static rule once inspections begin, and inspect even more frequently than that when negative evidence is accumulated. Utilize evidence from previous inspections to justify either restoration or another inspection. Cancel inspections and hesitate to restore the process at the end of a production run. Consider using scheduled restoration, in which restoration is carried out regardless of the results of any inspections. Implementation, limitations, and extensions are addressed.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 57 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.