Prioritizing customer orders in a high-end server manufacturing environment using artificial neural networks

Faisal Aqlan, Sarah S. Lam, Keila Martinez

Research output: Contribution to conferencePaperpeer-review


In today's competitive manufacturing environment, an integrated and dynamic production scheduling system would allow companies to respond quickly to customer orders and maintain their market leadership. Order scheduling is one of the most complicated tasks in high-end server manufacturing due to having customized orders that require quick turnaround time, in an environment that has limited resources. This environment drives the need for smart prioritization and scheduling of customer orders. The objective of order scheduling is to arrange processing sequence and starting and finishing times for orders to optimize certain criteria (e.g., makespan). This research uses an artificial neural network approach in conjunction with an analytic hierarchy process (AHP) to prioritize customer orders in a high-end server manufacturing environment. Orders can have different priorities based on factors such as order type, cycle time, and requested ship date. AHP is used to obtain the priority indexes for orders based on their attributes and the results from the AHP are used to train and validate the neural networks. Based on the predicted priority of the orders from neural networks, specific shipping dates are then assigned to orders while considering the availability of main commodities, orders currently in processing, and resource capacity.

Original languageEnglish (US)
Number of pages8
StatePublished - Jan 1 2012
Event62nd IIE Annual Conference and Expo 2012 - Orlando, FL, United States
Duration: May 19 2012May 23 2012


Other62nd IIE Annual Conference and Expo 2012
Country/TerritoryUnited States
CityOrlando, FL

All Science Journal Classification (ASJC) codes

  • Industrial and Manufacturing Engineering


Dive into the research topics of 'Prioritizing customer orders in a high-end server manufacturing environment using artificial neural networks'. Together they form a unique fingerprint.

Cite this