Improve Supply Chain Management Using Neural Networks and Regressive KPI Relationship Metamodels

The economic development of the emerging countries has been regarded in the recent time as a serious opportunity for cost reduction from western manufacturer, results obtained from this vision were the increase of the delocalization of the production process and the increase of the management complexity. In order to answer to the new market demand industry turn to software vendors looking for specific ERP systems(Davenport 1998) and starting specific projects for supporting Business Process Redesign (BPR). As seen in several industrial contexts few projects ended with success while the majority of them running very quickly out of budget and in serious delay.

In this sector authors identified a lack of anticipatory models able to drive the ERP implementation process to the right and they propose in this paper a meta modeling approach able to bridge this gap.

Proposed methodology integrates Data Analysis, Regression Meta-Modeling and Artificial Neural Networks processing in order to identiy hidden relationships among KPI and guide the BPR decision makers. The paper outline the proposed methodology as well as a practical application to a real life industrial case.