Behind the scenes @ Labs
cancel
Showing results for 
Search instead for 
Did you mean: 

HP Labs Middleware research enables Software-Defined Manufacturing

Guest_Blogger

Contributed by HP Labs Principal Staff Researcher Jun Zeng and HP Labs R&D Director Gary Dispoto

 

SDManuf-banner.jpg

 

Mass customization, in Tseng and Jiao’s definition, is "producing [customized, personalized] goods and services to meet individual customer's needs with near mass production efficiency".[1] It includes two seemingly competing objectives that make its realization challenging:

 

1) make individualized products (with a high value) to meet customers’ heterogeneous needs, resulting in ever smaller order sizes to the point where “every product is different” (EPID); and

 

2) deliver operational efficiency of a quality that successfully competes against mass production (think Henry Ford and assembly lines).

 

We are moving closer to widespread adoption of mass customization, however, thanks largely to technical innovations on two fronts. Firstly, we now have general-purpose machines that can produce diverse products (i.e. successive items of different shape and/or functionality) at more-or-less the same cost as creating copies of a single item in bulk (for example, via 3D printing). We also have software (middleware in particular) that can effectively and efficiently compose production workflows for individualized products and thread different machines (co-located, or geographically dispersed; of different capabilities, capacities, and availabilities) for workflow execution.

 

Our research at HP Labs aims to further the progress of mass customization in areas such as 3D printing. We’re aiming, in particular, to advance middleware research towards the creation of the data and compute fabric that could bring us a 3rd industrial revolution.

 

Middleware, for our purposes, is the software stack that lies between a product’s specification and the fulfillment system that creates it. It embraces translating customer orders into different candidate fulfillment workflows and selecting and executing a workflow to maximize operational efficiency under the constraints of service level agreements. It also includes the scheduling and routing of tasks to assigned machines,[2] monitoring machines (e.g., for material replenishment) to inform production management, and monitoring the fulfillment flow to inform upstream and downstream supply chain partners. Additionally, it needs to accommodate exceptions handling, such as when a particular production step fails a quality requirement, and up-to-the-moment cost accounting to inform product quotation. This middleware tier, in fact, is largely responsible for the agility, adaptability, and programmability in manufacturing – which has led us to coin the term Software-Defined Manufacturing to describe it, borrowing the phrase from IT industry[3]. This, we believe, is what will fundamentally enable a 3rd industrial revolution – from the effective use of 3D printing, to the ubiquitous deployment of industrial Internet of Things, to the big data analytics that will drive just-in-time decision-making.

 

Research into commercial and industrial digital printing has been at the forefront of mass customization. EPID-for-printing has been a reality for the past decade, from book-on-demand manufacturing to digital photo-book printing. Motivated by this foresight, five years ago we put together a collaborative research project to work with leading research institution Duke University and mass customization pioneer RPI.Together, we looked to frame manufacturing (in particular, mass customization) as a computing problem, and to develop algorithms and prototypes that would address the technical challenges outlined above. One result was an IEEE sponsored workshop, put together by HP Labs engineers with Rick Bellamy (RPI’s CEO) and Professor Krish Chakrabarty of Duke University, that shared our ideas with the broader academic and industrial community.

 

Book Cover.pngA new book, “Data-Driven Optimization and Knowledge Discovery for an Enterprise Information System” by Qing Duan (of Risk Data Sciences, Paypal Inc.) and Jun Zeng (HP Labs),[3] is another outcome of this effort. Based largely on Duan’s PhD dissertation, written under the guidance of Professor Chakrabarty, the book outlines the major research and industrial implications of the development of mass customization. Grounding her research in academic fields such as operations research, systems engineering, statistics, and machine learning, Duan provides a comprehensive review of the various relevant (and in particular, on-line) mass customization optimization and prediction techniques. In addition, she identifies the key challenges facing mass customization and presents significant algorithms and prototypes that have emerged from leading-edge research in this domain. The book covers a wide range of topics, from task scheduling and resource allocation to workflow optimization, process time and status prediction, order admission policies optimization, and service-level performance analysis and prediction. It also provides an in-depth look at advances in automation and reconfigurability-based fault tolerance, and explores how to obtain data-driven recommendations for effective decision-making. Overall, it’s an excellent introduction to, and overview of, the subject – useful as a technical textbook for graduate level operational research and management science, and as a valuable technical reference for industrial researchers and practitioners alike.

 

[1] Tseng, M.M.; Jiao, J. (2001). Mass Customization, in: Handbook of Industrial Engineering, Technology and Operation Management (3rd ed.). New York, NY: Wiley. ISBN 0-471-33057-4.

[2] This may also include, based on specific machines, translating the tasks into machine instructions, e.g., G-Codes. This is where the middleware and the machine firmware blend. We have ongoing research on this front for 3D printing.  

[3] Other industrial practitioners are adopting similar phases, e.g., software-defined supply chain, http://www-935.ibm.com/services/multimedia/The_new_software-defined_supply_chain_Exec_Report.pdf

 

 

 

0 Kudos
About the Author

Guest_Blogger

Events
June 18 - 20
Las Vegas, NV
HPE Discover 2019 Las Vegas
Learn about all things Discover 2019 in  Las Vegas, Nevada, June 18-20, 2019
Read more
Read for dates
HPE at 2019 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2019.
Read more
View all