Supply chain monitoring: a statistical approach.

MELE, FERNANDO D.; MUSULIN, ESTANISLAO; PUIGJANER, LUIS

15th European Symposium on Computer Aided Process Engineering (ESCAPE-15)

Elsevier

Lugar: Amsterdam; Año: 2005; p. 1375 - 1380

Although the nodes of a supply chain (SC) network generate a huge amount of data along their operation, extracting useful information from them is not straightforward. Within the Supply Chain Management (SCM) scope, monitoring reveals as a key task that is currently waiting for further study. It is necessary to minimize risks of undesired situations and administrative efforts to manage material flows. Supply Chain Monitoring (SCMo) techniques should support manager decisions warning of the abnormal situation telling what have gone wrong and suggesting solutions. Additionally, they should be able to store the causes and consequences in order to help in the decision making onto future similar situations. This work presents an extension of multivariate statistical methods to SCMo that consists in a wavelet based multi-scale Principal Component Analysis (PCA) technique accounting for time delays. The proposed approach has been tested using data generated through an event discrete simulation model running in several scenarios. Results have revealed that statistical multivariate techniques are very useful for SCMo.