In search of optimality: A re-cast

I’ve renamed my blog to ‘in search of optimality’ from ‘in search of excellence’.

My recent reconnection with systems modelling reminded me that many, many years ago I wrote a report ‘In search of optimality: a systems technologist goes east’ (Extracts summarised in Mellalieu, 1985). The report detailed the interesting learnings I had identified based on my six months post-doctoral study tour based at Massachusetts Institute of Technology (MIT) and Lancaster University, sponsored by the New Zealand State Services Commission (NZSSC) and NZ Department opf Scientific and Industrial Research (DSIR). In both locations, I wasted hosted by the departments of operational research. My external examiner for my PhD was also from Lancaster, and I also met Peter Checkland, of soft systems fame.

The title of my report is an allusion to Peters and Waterman’s book ‘In search of excellence’. Their book had been published whilst I was in America. I read the book with great excitement,noting their examples of business excellence around me in the many places I visited.

I had  completed my PhD in which I made extensive use of applied optimisation theory. So the title of my report ‘In search of optimality’ came naturally. It now seems natural that my web-site should thus be renamed.

So what is optimality? Optimality occurs when you find the best available values of some objective function given a defined domain (Wikipedia, optimisation).I’ll explain….

During my PhD, my search was for the best way to reconfigure the domain of the New Zealand dairy industry to maximise long-term economic performance. My objective function incorporated:

• expected future values of dairy products
• factory processing costs and fixed costs
• factory process yields of products and by-products
• transport costs of milk from farm to factories, and by-products to by-product processing facilities

The problem was ‘constrained’ by factors including: factory capacities, tanker capacities, product demand, and the quantity of milk produced by farms. These factors varied on a month-by-month basis. Furthermore, there were significant seasonal and long-term variations to be considered. Mellalieu & Hall (1983)

More generally, my search engaged me in formulating the problem in mathematical terms: as a mathematical model. Once formulated as a mathematical model, mathematical algorithms can be brought to bear to find the optimal solution. In school, for example, we learn to use differential calculus to find optimal points on curved spaces. Example: A stone is thrown upwards from the surface of the moon with gravity g/6 with speed v at an angle a. What is the maximum (optimal) height that the stone reaches. What distance, d from the thrower does the stone land? Knowing that the stone travels in a parabolic curve, we can solve this problem using calculus. (For a complicated (!!!) answer to this question, see: Parabola Separation Queries and their Application to Stone Throwing, Otfried Cheong1, Hazel Everett2, Hyo-Sil Kim1 , Sylvain Lazard2, & Ren´e Schott)

More generally, there are a host of procedures that one might choose, depending on the nature of the mathematical formulation used to represent the problem under investigation. Some procedures use calculus. Others use cleverly-conceived algorithms such as Danzig’s linear programming (LP). Many procedures can be executed with pencil and paper - such as those involving the principles of calculus. However, most real-world problems require computers to manage the data sets and execute the algorithms required to identify optimality.

Part of the challenge in optimisation studies (a branch of operations research) is to identify what factors (or variables) should be incorporated into the model. How far into the future should one consider? Beyond readily-measured financial costs and market prices, how does one quantify factors such as the impact of ‘environmental footprint’? This latter problem is one of ‘multi-objective optimisation’.

Is excellence is a subset of optimality? Or are they a different names for the same concept?

If you are not searching for optimality, then what are you doing?

Postscript

A good explanation of the stone throwing problem is illustrated by the physics of the medieval/Chinese trebuchet machine. The physics, of course, is also relevant to catapaults, ballistas, guns, and rockets.

Physics of the Trebuchet. (2010, October 23). library.thinkquest.org. Retrieved October 23, 2010, from http://library.thinkquest.org/05aug/00627/phy.html

Reference

Mellalieu, P. J. (1985). Some New Directions in Systems Modeling Practice. New Jealand Journal of Technology, 1(4), 223-238.   

Mellalieu, P. J., & Hall, K. R. (1983). An Interactive Planning Model for the New Zealand Dairy Industry. Journal of the Operational Research Society, 34, 521-532. doi:10.1057/jors.1983.119 

Related articles

• INVITED BLOGGER OF THE MONTH: On choosing to teach: a professor reflects (Part 1) (ictelt.blogspot.com)
• Dr. Vladimir A. Masch: Harnessing Our Technological Strengths for the Economy: Risk-Constrained Optimization (huffingtonpost.com)
• SAS Marketing Optimization Produces ROI for Commerzbank (eon.businesswire.com)
• Core plus mathematics-overview (slideshare.net)
• Vistaar Expands Optimization Expertise with David Glenn (prweb.com)
• Jeff Sutherland, Co-Creator of Scrum and qaSignature Announce Partnership to Accelerate Agile Development and Testing (eon.businesswire.com)
• KSS Fuels to Present Session on Fuels Price Optimization at EURO XXIV Conference (eon.businesswire.com)
• Carnegie Mellon researchers break speed barrier in solving important class of linear systems (eurekalert.org)
• Pioneering Mathematician László Lovász to Receive Inamori Foundation’s 26th Annual Kyoto Prize for Lifetime Achievement in “Basic Sciences” (eon.businesswire.com)
• Risk Solver Platform for Excel Uses Windows HPC Server 2008 R2 to Solve Challenging Optimization and Simulation Problems on Compute Clusters (prweb.com)
• Artificial Intelligence | Planning and Scheduling Software (planning-and-scheduling-software.org)
• Mathematics and the environment (eurekalert.org)
• Human intelligence and machine stupidity: Supply chains are about effectiveness, not only efficiency (kinaxis.com)
• Learnings from In Search of Excellence (christopherscottblog.typepad.com)
• The Equation Of Excellence (owengreaves.com)

I am a management scientist! Re-optimising a career

Image via Wikipedia

Following a chance meeting with a distant friend and former colleague, I’ve been re-connecting with my former life as a management scientist a.k.a. operational researcher. Accordingly, I’m reviewing some ‘found object’ textbooks in my library. I particular, I am enjoying reading Maani and Cavana’s (2000) introductory text on systems thinking and systems dynamics. It’s rather a refreshing task: rather like listening to a Bach Cantata.

My most recent engagement with the MS/OR domain has been my ten semester’s teaching of competitive business strategy using a computer-based ‘learning laboratory’ (micro-world). The learning laboratory utilises the Business Strategy Game (BSG) developed by Thompson et al. I have also developed a few simple financial planning spreadsheets for my teaching of new venture business planning.

I was first introduced to computer-based business strategy games by my physics teacher whilst a pupil at Cambridge High School in 1971. I continued the experience as a student of industrial technology, then constructed my own strategy ‘game’ during my doctoral studies. However, that game was more seriously called a a Decision Support System for long-range planning (Mellalieu, 1982; Mellalieu & Hall, 1983).

I realised late last year that there was an opportunity to complete some ‘unfinished business’ from my doctoral thesis. Consequently, I’ve been updating my analysis of the issue of ‘smoothing the flush’ in New Zealand dairy milk processing (Mellalieu, 2010)

I’ve taken a break from teaching my unique BSG-based strategic thinking course. As I immerse myself in upgrading my knowledge of new opportunities in systems modelling, decision support systems, exploratory data analysis, genetic algorithms, and data mining, I am spotting opportunities for completing an exciting re-creation of the course.

Life-changing event

The following is the conference that changed my life for almost a decade. I submitted my doctoral research proposal to Troughton and Hall, who promptly recruited me to join their biosystems modelling group at the Physics and Engineering Laboratory (PEL), now Industrial Research Ltd. Later, all three of the team re-located to the Ministry of Agriculture (MAF), where we established a systems modelling capacity for industry policy analysis. Kearney was a later member to join the team. Later, he was recruited to join the SAS Institute (US) to continue developing his ultra-powerful network optimisation algorithms. One special attraction to me for joining the DSIR PEL team was that my doctoral research would assist directly in improving industrial productivity of key economic sectors.

Troughton, J., & Hall, K. R. (Eds.). (1977). Management of dynamic systems in agriculture. Proceedings of Management of dynamic systems in agriculture, Gracefield, Lower Hutt, NZ: Department of Scientific and Industrial Research (DSIR).

The aftermath: Troughton, Hall, Mellalieu and the biosystems ‘fair-children’

AgResearch Professorial Chair in Systems Thinking | Scoop News. (2008, September 24). . Retrieved October 11, 2010, from http://www.scoop.co.nz/stories/SC0809/S00072.htm

Introduction to Optimization: PROC NETFLOW. (n.d.). . Retrieved July 5, 2009, from http://support.sas.com/documentation/cdl/en/ormpug/59679/HTML/default/intromp_sect8.htm [Kearney’s magnificent algorithm in user-friendly form!]

Kearney, T. D. (1999). Advances in Mathematical Programming and Optimization in the SAS System. SUGI Proceedings, SAS Institute. Retrieved from http://support.sas.com/rnd/app/papers/mathprog.pdf

Kearney, T. D., Hall, K. R., & Mellalieu, P. J. (1984). Recent Advances in Network Optimization Methods and Applications. In Proceedings of the Annual Conference of the United Kingdom Operational Research Society. Presented at the Annual Conference of the United Kingdom Operational Research Society.

Mellalieu, P. J., & Emerson, A. (2009). Developing reflective learning in a strategic thinking class. In Unitec Teaching and Learning Symposium. Presented at the Unitec Teaching and Learning Symposium, 28 September 2009, Auckland, NZ: Unitec Institute of Technology. Retrieved from http://web.mac.com/petermellalieu/Teacher/Blog/Entries/2009/9/29_Symposium%3A_Developing_reflective_learning_in_a_strategic_thinking_course.html

Mellalieu, P. J., & Emerson, A. (n.d.). Course: Strategy and Enterprise Thinking - Announcement APMG 6340. Unitec Business School Publications. Retrieved July 11, 2009, from http://web.mac.com/petermellalieu/UBSpublications/BizBits/Entries/2009/7/11_Course:_Strategy_and_Enterprise_Thinking_-_Announcement_APMG_6340.html

Mellalieu, P. J., & Turner, K. D. (1985). Expert Systems for Agricultural Production [crop damage in agricultural crops]. In Proceedings of the Operational Research Society of New Zealand.

Mellalieu, P. J. (1977). Management aids in the biological industries (Conference summary of DSIR Management of dynamic systems in agriculture). Productivity & Technology, NZ Department of Trade & Industry, (6).  

Mellalieu, P. J. (1982). A Decision Support System for Corporate Planning in the New Zealand Dairy Industry (Doctor of Philosophy in mathematics, statistics and operations research). Victoria University of Wellington. Retrieved from http://hdl.handle.net/10063/568  

Mellalieu, P. J. (1985). Some New Directions in Systems Modeling Practice. New Jealand Journal of Technology, 1(4), 223-238.  

Mellalieu, P. J. (2010). Responding to the threat of price volatility in land-based industries: Economic impacts of ‘smoothing the flush’ in New Zealand dairy milk production and processing. Presented at the Unitec Learning, Teaching, and Research Symposium, Auckland, NZ: Unitec Institute of Technology. Retrieved from http://web.me.com/petermellalieu/Teacher/Examples/Entries/2010/9/28_smoothing_the_flush_in_New_Zealand_dairy_milk_production_and_processing.html

Mellalieu, P. J., & Hall, K. R. (1983). An Interactive Planning Model for the New Zealand Dairy Industry. Journal of the Operational Research Society, 34, 521-532. doi:10.1057/jors.1983.119  

SAS/GRAPH(R): Network Visualization. (n.d.). . Retrieved July 5, 2009, from http://support.sas.com/documentation/cdl/en/grnvwug/61307/HTML/default/n1a2kt6ftdrjs6n10ziku07mz73d.htm [See Mellalieu and Hall 1983 for the pre-cursors to this procedure]

References

Decision support system - Wikipedia, the free encyclopedia. (n.d.). . Retrieved February 3, 2010, from http://en.wikipedia.org/wiki/Decision_support_system

Maani, K. E., & Cavana, R. Y. (2000). Systems Thinking and Modelling: Understanding Change and Complexity (1st ed.). Pearson Education New Zealand.  

Marakas, G. M. (2002). Decision Support Systems in the twenty-first century: DSS and data mining technologies for tomorrow’s manager (2nd ed.). Upper Saddle River, NJ: Prentice Hall. Retrieved from www.prenhall.com/marakas

Thompson, A. A., Stappenbeck, G. J., Reidenbach, M. A., Thrasher, I. F., & Harms, C. C. (n.d.). Business Strategy Game Simulation. Retrieved July 7, 2009, from http://www.bsg-online.com/

Related articles

• SAS Optimization Software Saves Hub Group $11 million - 2200% ROI (eon.businesswire.com)
• An Introduction to General Systems Thinking- Weinberg [Book] (dorsethouse.com)
• Management Education (cognitive-edge.com)