Yesterday was one of those remarkable days when everything seems just about right. The highlight was an email message I received from a Prof. Fang from Ryerson University notifying us that we had been both nominated and awarded the Omond Solandt Award by the Canadian Operational Research Society for ongoing and outstanding contribution to the field of Operations Research (OR). No, it’s not a Nobel Prize or an Oscar, but whenever a group of smart people publically recognize our work, the honor and pride are genuine.
What is even more special is the fact that the OR often seemed to be a world in itself and perhaps a little distant from the core symbolic computation or engineering modeling communities. In a nutshell, OR develops complex models to manage logistics (i.e. shipping, transportation, networks, etc.). Modern origins of this field date back to World War II, though some claim that the famed Charles Babbage himself was the founder when he applied principles of costing in the development of the British “Penny” Post system.
Twenty years ago, the computational subfields of OR were definitely distant cousins to our immediate domain of symbolic computation. Although both were involved in some form of mathematical modeling, OR modeling required highly specialized “mathematical programming” algorithms capable of dealing with tens of thousands of variables. And at that time, Maple’s capabilities in this field were far from state of the art.
I’m happy to report that today, this is no longer the case … as indicated by the award in many ways. When you think of it, it makes perfect sense. Modern Maple has been richly empowered, over the past decade, with new algorithms for linear algebra, graph theory and other combinatorics, statistics, and visualization. These are the fundamental computational primitives of OR and Maple can hold its own on the basics. Heck, we even introduced a fully capable Global Optimization Toolbox (GOT).
The future is even more intriguing. In the context of the physical modeling applications, one of the proverbial “Holy Grails” is optimal design and optimal parameter management. The growing complexity of modern engineering systems means there are many more parameters and factors to analyze and manage during the modeling and simulation process. Indeed the core capabilities of symbolic computation have been identified as a key technology in helping us efficiently manage parameters. And today’s techniques, though a great start, have not realized the potential of true optimal design. Of course, there are specialized case studies where we’ve successfully achieved optimal designs for well defined systems, but the expansion of these techniques into broadly accessible techniques by industry at large is still more a vision than reality.
Optimal design is fully automated intelligent design strategies where all important engineering variables and parameters are simultaneously and algorithmically balanced to yield the best possible design with the minimal iterations. This fancy mathematical juggling act is really what the OR community is particularly good at. Whether it’s cost-effectively shipping tropical fruit to my town’s freezing winter climate or balancing fuel efficiencies with the countless performance and financial cost parameters of a new engines, the convergence of the best of OR with the emerging needs of engineering design is not only welcome, but essential to the productive evolution of design methods for the future.
The Omond Solandt Award is certainly an honor, but more importantly, it’s a yet another indicator our scientists and engineers are continuing to construct collaborative bridges among fields that have historically been separate. Realizing true optimal design technology is a formidable task but it heartens me that we continue to take concrete steps forward.