A report on the development of the control chart

Contardo, Ivona (2007-03)

Thesis (MComm) -- University of Stellenbosch, 2007.

Thesis

ENGLISH ABSTRACT: The goal of quality control as stated by Feigenbaum (1961) is to provide a product or service into which quality Is designed, built, marketed and maintained at the lowest economical cost which simultaneously allows for full customer satisfaction. Statistical process control techniques, specifically control charts, are widely employed to achieve this goal. Walter A. Shewhart developed the control chart in 1924 in order to differentiate between random causes of variation and assignable causes of variation. In situations where assignable causes occur, a corrective action should be taken to return the process to the in-control state. A process should be able to operate in the in-control state for a relatively long period before an assignable cause will come about. The heuristic design of Shewhart however, was not guaranteed to be economically optimal. In 1956 Duncan proposed that the design parameters of control charts should be chosen in a manner that minimises the economic costs associated. Since then various developments in the economic design of control charts have taken place. Some research has been done to increase the power of the control chart and consequently the statistical design of control charts has been presented. The statistical design of control charts is designed in a manner as to place constraints on the control chart. Saniga in ( 1989) proposed to combine the economic and statistical designs in the economic statistical design in an attempt to minimise the economic costs of the control charts under some statistical constraints. Results obtained by various authors show that the economic statistical designs perform best in the sense of achieving the desired statistical properties while simultaneously minimising the associated costs. Various cost models have been developed under different distributions and the expressions for the expected cycle time and expected cycle cost have been derived. In 2005 Yang and Rahim developed a cost model for the economic statistical design of control charts for a process with multiple quality characteristics under a Weibull shock model. The development of this model will be discussed in detail.

AFRIKAANSE OPSOMMING: Die doel van kwaliteitsbeheer soos voorgestel deur Feigenbaum (1961) is om 'n produk of diens te lewer wat van so aard is dat kwaliteit in die produk se ontwerp, samestelling, bemarking ingesluit is en volgehou word teen die laagste ekonomiese koste wat steeds voile verbruikers tevredenheid verseker. Statistiese kwaliteitsbeheer tegnieke, met spesifieke verwysing na kontrole kaarte, word ingespan ten einde hierdie doel te bereik. Walter A. Shewhart het in 1924 die kontrole kaart ontwikkel met die doel om tussen kans oorsake en aanwysbare oorsake van variasie te onderskei. In omstandighede waar 'n aanwysbare oorsaak teenwoordig is, word 'n korrektiewe aksie vereis om die proses terug te bring na die in-kontrole toestand. 'n Proses behoort vir 'n relatiewe lang periode in die inkontrole toestand te funksioneer voor die voorkoms van 'n aanwysbare oorsaak. Die heuristiese ontwerp van Shewhart waarborg egter nie dat die kontrole kaart ekonomies optimaal sal funksioneer nie. In 1956 stel Duncan voor dat die seleksie van parameters vir die ontwerp van kontrole kaarte op so manier geskied dat dit die ekonomiese koste minimeer. Sedertdien is verskeie verbeterings in die ekonomiese ontwerp van kontrole kaarte gemaak. Navorsing is ook gedoen ten einde die onderskeidingsvermoe van die kontrole kaart te verbeter en sodoende is die statistiese ontwerp van kontrole kaarte voorgestel. Die statistiese ontwerp plaas statistiese beperkings op die kontrole kaart. In 1989 stel Saniga voor dat die ekonomiese en statistiese ontwerpe gekombineer word om die ekonomiese statistiese ontwerp te vorm. Die doel van die ekonomiese statistiese ontwerp is om die ekonomiese koste te minimeer gegewe sekere statistiese beperkings. Resultate toon dat die ekonomiese statistiese ontwerp beter presteer en die statistiese beperkinge kan hierdeur behaal word terwyl die toepaslike kostes terselftertyd geminimeer word. Verskeie koste modelle is ontwikkel ender verskillende verdelings en uitdrukkings vir die verwagte siklus tyd en verwagte siklus koste is afgelei. In 2005 ontwikkel Yang and Rahim 'n koste model vir die ekonomies statistiese ontwerp van kontrole kaarte vir 'n proses met meerveranderlike kwaliteitseienskappe ender 'n Weibull skok model. Die ontwikkeling van hierdie spesifieke model sal in besonderhede bespreek word.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/50683
This item appears in the following collections: