Planned spontaneity for better product availability

Abstract

This invited article reflects the particular preoccupations of the author, which concern the parallel and connected development of academic models for managing inventories and their actual applications in industry.

Some centralised production-inventory models that have been of service to contemporary business practices in decentralised decision-making settings are discussed here. I have limited my examples to those in which I have had direct experience. I have attempted to highlight four key points. (a) The strategic role of inventories in capturing revenue and market share, in addition to their traditional role as buffers, in our contemporary customer-scarce Schumpeter’s market. (b) Facilitating the acceptance of model outputs by decision makers across organisational silos and even across firm boundaries using management mechanics. (c) Development and use of a general-purpose algorithm utilising infinitesimal perturbation analysis (IPA) derivatives. (d) The emergence of enterprise inventory optimisation (EIO) as mainstream software.

Keywords:

• inventory management
• operations management
• operations strategy
• supply chain design
• safety stocks

Acknowledgements

There are so many people who have helped make inventory models useful to practice – ranging from my PhD students, faculty colleagues, co-authors and collaborators, SmartOps’ employees and channel partners, executives, IT staff and planners from Fortune 500/Global 2000 firms – that to list them individually by name would exceed any reasonable page limit. I would like to thank Alan Scheller-Wolf, Mustafa Akan, Tinglong Dai, Steve Graves, Jack Muckstadt, Soo-Haeng Cho, Mike Trick, Nicola Secomandi, Jay Swaminathan, Paul Glasserman and David Simchi-Levi for their help and suggestions in the preparation of this manuscript.

Notes

1. It is said that imitation is the sincerest form of flattery. The style of the article – especially the extensive use of quotes – pays homage to the thoughtful articles of S. Chandrasekhar collected in Truth and Beauty: Aesthetics and Motivations in Science (1987). The use of some language in the article is borrowed from the preface of Tjalling C. Koopmans (1957), Three essays on the state of Economic Science, while some language in the concluding section is from Milton Freidman, from his conclusion section of a 1955 report found in Two Lucky People (Friedman and Friedman 1998). I also benefited from reading the personal narratives in the collection of articles in the fiftieth anniversary issue of Operations Research (January–February 2002) and of Management Science (December 2004).

2. The desire for academics to help senior management solve their most pressing problems is not new. James O’ McKinsey, a professor at University of Chicago Graduate School of Business (now called Booth School of Business) founded McKinsey & Company in 1926. He wanted to do more than what efficiency experts were doing under the banner of management engineering. Marvin Bower later coined the term ‘management consulting’ (see http://www.mckinsey.com/about_us/history/1930s) which has remained in use since the 1930s. You should now understand why I prefer ‘management mechanics.’

3. The earliest reference I could find that seems to have put F.W. Taylor on the path towards Scientific Management, the precursor to management engineering – and looking at management from a scientific perspective, leading to the creation of the fields of Industrial Engineering, Industrial Administration and Operations Management – is a 1886 paper by Henry R. Towne ‘The Engineer as an Economist.’ Towne was the president of the Yale and Towne Manufacturing Company and a president of the American Society of Mechanical Engineers (ASME). The young Taylor (who became an ASME member in 1885), impatient with what he called ‘just management of initiative and incentive’, wanted to think about how to ‘increase productivity and lighten labor’s efforts.’ (Taylor 1911)

4. The phrases ‘production scarce’ and ‘customer scarce’ are borrowed from Yuji Ijiri’s chapter in The Innovative University (2004).

5. My PhD advisor’s PhD advisor – my academic grandfather, so to say – is Arthur F. Veinott, Jr, whose academic grandfather is none other than T.E. Harris! I discovered this through the Mathematical Genealogy Project.

6. One of my minors in graduate study was queuing theory and N.U. Prabhu was on my thesis committee. The other minor was operations management, and Dick Conway was also a committee member. My PhD advisor was Robin Roundy. My PhD thesis was on stochastic models of serial production lines operated by kanban systems; see Tayur (1992, 1993b) and Muckstadt and Tayur (1995a, 1995b). The stochastic cyclic schedules in Tayur (2000) at GE are operated using kanban cards. My first introduction to inventory models was in a class taught by Jack Muckstadt at Cornell using Hadley and Whitin as a reference text. Bill Maxwell taught a course on scheduling. My first OR course was taught by T.T. Narendran at Indian Institute of Technology, Madras. Moving in the other direction, Nihat Altintas, Carlos Bispo, Feryal Erhun, Srinagesh Gavirneni, Roman Kapuscinski, Pinar Keskinocak and Jay Swaminathan are some of my PhD students.

7. I met Paul Glasserman in 1988 at Bell Labs in Holmdel, NJ. Beyond IPA, Paul and I have developed a large deviation approximation (based on extreme value theory) for capacitated multi-echelon inventory models (Glasserman and Tayur 1996; Glasserman 1999). A single-stage formula connects inventory levels, service levels, excess capacity and variability in closed form. This Glasserman–Tayur formula (for discrete time, capacitated inventory model) can be considered a cousin of Pollaczek–Khinchin (M/G/1 queues).

8. Rekha Thomas was a PhD student at Cornell and her PhD thesis is on Grobner basis and integer programmes. Dimitris Bertsimas and Georgia Perakis were hosts during my 1997 sabbatical at MIT.

9. Holt and Simon worked closely with PPG. Magee (and Arthur D. Little) worked closely with Johnson & Johnson. Both PPG and Johnson & Johnson have integrated SmartOps EIO software into their IT stacks, connected it to Oracle and SAP ERP and APS, and have made multi-stage inventory planning part of their monthly global SIOP process.

10. I have limited my discussion to four or so PPOMs in this article. Just as grocery retailers are optimising NLC, their suppliers (like Heinz) are optimising quantity discount schedules (see Altintas, Erhun, and Tayur 2008) and, as ConAgra Foods, are making their own production planning more flexible while reducing total costs (see Mehrotra et al. 2011).

11. I have not discussed the role of information (Gavirneni, Kapuscinski, and Tayur 1999), limited history of data (Akcay, Biller, and Tayur 2011), operations reversal as a strategy to manage variety (Kapuscinski and Tayur 1999), guaranteed lead times (Kapuscinski and Tayur 2007; Keskinocak, Ravi, and Tayur 2001) international operations affected by exchange rates (Scheller-Wolf and Tayur 2009), real (capacity) options (Erhun, Keskinocak, and Tayur 2008a), supply chain co-ordination (Erhun, Keskinocak, and Tayur 2008b) or the role of the internet and e-business (Keskinocak and Tayur 2003; Swaminathan and Tayur 2004).

12. I have also not discussed the choice of the product portfolio itself. See Yunes et al. (2007) for what Deere has. Deere has also tailored its distribution logistics by season to further improve operations performance with respect to responsiveness and cost; see Tardif et al. (2010).

13. In 2005, we tested a massively parallel version of our algorithm on IBM Blue Gene (using their on-demand Deep Computing offering). An industrial instance with about a million inventory targets – one for each item, location and week – solved in 0.04 s on a half-rack system of 512 parallel processors.

14. I have discussed monetising the value of operations research in Camm and Tayur (2010). I presented EIO (and entrepreneurship) at UCLA’s Marschak Colloquium, along with private equity and lean operations (January 2012).

15. Entrepreneurship and significant contributions to practice (in diverse areas) by OR/OM professors have a long history: Harry Markowitz, Egon Balas, Art Geoffrion, Jack Muckstadt, Don Ratliff, Marshall Fisher, Bob Bixby, Morris Cohen, Steve Graves, Sunder Kekre, Hau Lee, David Simchi-Levi, Larry Wein and Dimitris Bertsimas, to name a few. notable people