SGER: Multiscale Methods for Supply Chain Monitoring

This Small Grant for Exploratory Research (SGER) provides funding for the exploration of multiscale methods for monitoring the timely and correct movement of products and transactions through supply chains, that is, from their originating order through their delivery to the customer. The methods will use individual product status information, which can be provided by radio-frequency identification (RFID) or other computerized tracking technology. Each data element will be viewed as a triple of the product ID, the status/location of the product, and the time of the status check. The aggregation of these triples over time, ID, and/or status provide ways to view the data on different scales, from the low level detail of a part moving though a manufacturing plant to the high level status of an order consisting of a group of products as it moves from order to delivery to the customer. The aggregated and disaggregated data will be examined using specialized multivariate statistical process control (SPC) methods. Existing multivariate SPC methods will be enhanced to allow the capture of dependencies in product movement that come from the structure of the supply chain network, using clustering and network optimization algorithms. The methods will be tested against supply chain data provided by corporate members of the Center for Supply Chain Research at Penn State. The center's corporate members include Dell, Penske Logistics, Hershey Foods and Johnson & Johnson. The results of this work will be disseminated through Center meetings, conference presentations, and the curriculum of the Department of Supply Chain and Information Systems.

If successful, the results of this research will lead to more effective ways to detect improper routing and delivery delays in complex supply chains. The multiscale character of the methods will allow them to be used to manage the supply chain at both tactical and strategic levels, and to provide quick detection and root-cause diagnosis of supply chain problems. The multiscale methods will contribute to general knowledge about multivariate SPC, a widely used Six Sigma tool.