Thesis (MScIngwet)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: This thesis is a case study in optimal inventory control, applied to Clickabox factory, a South African cardboard box producer from whom cardboard boxes may be ordered at short notice via the internet. The problem of developing a decision–support system for optimal stockholding at the factory, in order to minimize cardboard off–cut wastage subject to required service levels, is addressed in this thesis. Previously a simple replenishment policy, based largely on experience, was implemented at the factory. The inventory model developed for and applied to Clickabox in this thesis takes account of a raw materials substitution cascade, as well as the stochasticity of demand, and other factors such as cost, service level and spatial requirements for the storage of stock. This combination of stochastic demand and product substitution has not, to the author’s knowledge, previously been dealt with in the literature. There are two primary deliverables of this study. The first is a suggestion as to the suitable stock composition (cardboard types from which boxes may be manufactured) to be kept in inventory at the factory. The second deliverable is a computerised decision–support system, based on the inventory model developed, to aid in future inventory replenishment decisions at Clickabox. Some of the results of this thesis have, at the time of writing, already been implemented with success at the factory. These include the suggestions given to the management of Clickabox as to the suitable stock types to be held in inventory, which have been implemented in stages since March 2003. The suggested stock composition has proven to be superior to the previous stock types held, in terms of a reduction in off–cut wastage and increased availability of suitable boards. / AFRIKAANSE OPSOMMING: Hierdie tesis is ’n gevallestudie in optimale voorraadbeheer, toegepas op Clickabox fabriek, ’n Suid–Afrikaanse kartondoosprodusent by wie kartondose op kort kennisgewing via die internet bestel kan word. In hierdie tesis word ’n besluitnemingsteunstelsel ontwikkel vir optimale bestuur van voorraad by die fabriek, wat karton afknipselvermorsing onderhewig aan vereiste diensvlakke minimeer. Vantevore is ’n eenvoudige voorraad aanvullingstrategie, wat hoofsaaklik op ondervinding gebaseer was, by die fabriek toegepas. ’n Wetenskaplike gefundeerde voorraadmodel word vir Clickabox ontwikkel en toegepas, waarin ’n rou–voorraad kaskade– substitusie proses in aanmerking geneem word, asook die stogastiese vraag na kartondose en faktore soos prys, diensvlakke en benodigde stoorruimte. Hierdie kombinasie van stogastiese vraag en rou–voorraad kaskade–substitusie is, tot die skrywer se kennis, nog nie in die literatuur behandel nie. Die studie het twee hoof–uitkomste ten doel. Die eerste is ’n aanbeveling ten opsigte van ’n geskikte rou–voorraad samestelling (kartontipes waaruit kartondose geproduseer kan word) wat by die fabriek in voorraad gehou moet word. Die tweede is ’n rekenaarmatige besluitnemingsteunstelsel, wat op die ontwikkelde voorraadbeheermodel gegrond is, en wat vir toekomstige besluite in verband met voorraadaanvulling by Clickabox bedoel is. Van die resultate wat in hierdie tesis vervat is, is reeds ten tyde van die opskryf daarvan doeltreffend by die fabriek ge¨ımplementeer. Ondermeer is die aanbeveling in verband met die geskikte voorraadsamestelling, geleidelik vanaf Maart 2003 by die fabriek ingefaseer. Dit het duidelik geword dat hierdie samestelling beter as die vorige voorraadprofiel funksioneer, in terme van ’n verlaging in afknipselvermorsing en ’n verhoging in die beskikbaarheid van geskikte kartonne.