Makers' Missing-Link: OmniQuest Optimization

And The Inspiration for Automating Automation

38 years ago this past January, my older brother Bernie asked our college roommate and me to help with his Electrical Engineering Senior project for The Ohio State University College of Engineering. Bernie and I, recently returned from internships programming robots at Cincinnati Milacron and IBM, respectively, found the use of these far more complicated than necessary.

We agreed Machines should simply be ‘walked through the work’ by the Human and, through the experience, automatically create an executable instruction set. 

Learning by Doing (even better: Learning by Teaching) became an obsession.

We pulled it off – using Bernie's Armatron toy robot, a block of wood, jig-saw and drill + dozen solenoids + pulleys + waxed cotton cord, a circuit board patched into a Radio Shack TRS-80 – and some slick BASIC programming. Bernie won 1st place in the The Ohio State University College of Engineering contest – leading to his landing a job in the GE FAST - Factory Automation Systems Technology program.

The Armatron began Bern's and my exposure to Machine-Based ‘Experiential Learning’ – teaching machines to learn; Automating Automation. A decade after Bernie's departure from this life I continue in the pursuit of enabling constellations of machines to self-organize around work - this means enabling computers to program themselves and program other computers. The future is fast approaching where computers do what we Need and Want without Humans mindlessly staring at refracted light and poking them.

Three key ingredients make this practical and self-sustainable: #1 Physical Sciences, #2 Management Science, and ? #1 has been addressed, for example, by Mark's Standard Handbook for Mechanical Engineers, while #2 is addressed by Plossl & Wight's Production and Inventory Control – Principles and Techniques, and #3 is BLANK.....until (for at least this Maker) now.

Through this long-running pursuit, I have had the great fortune to cross paths with an extraordinary number of individuals who have positively influenced the ascendance of Humanity through the creation of Tools and the assembly of Tools into Machines that leverage the capability of the Human to accomplish Work. Until I had the privilege of working for OmniQuest, this yet-to-be defined and ever-elusive third ingredient remained unknown to me.

I pause and reflect, as a Braker should do (Note: See "Third Side of the Coin - Explaining Maker-Braker/Braker-Taker" LinkedIn article previously published). My life's journey has been blessed with meeting one of the pioneers; likely the greatest single contributor of creating, evangelizing, and advancing the field of Optimization: Dr. Garrett N. Vanderplaats. His life's work is a Missing Link - the third ingredient: #3 Optimization.

With Optimization, the Maker is able to create a closed-loop feedback way of automatically adjusting for variance of the Actual from a Planned outcome. Dr. Vanderplaats' Multidiscipline Design Optimization, preceded by Numerical Optimization Techniques For Engineering Design, complete the major ingredients - provides the 'Missing Link' - demanded of Automating Automation. And the journey continues.

OmniQuest offerings implement Dr. Vanderplaats' Optimization Engines which serve to complete the major ingredients - providing the 'Missing Link' - demanded of Automating Automation.  

"What does this mean?" - the banner image of this article is that of an agricultural aerial drone tending a rice field. This Machine is programmed to perform its work based on a set of parameters the Human must maintain such as: Speed, Water/Fertilizer Flow Rate, Altitude and the relative position - Longitude and Latitude - from its starting point.

When one of the 'parameters', such as wind speed unexpectedly changes (e.g. a big gust of wind followed by an increase in the sustained wind, barometric pressure falling, temperature changing) the Machine must update its Plan. Optimization enables monitoring of all the Inputs (e.g. parameters) and Outputs so that course-corrections may be determined with the best possible outcome on a continuous basis.

Consider a fleet of 100s of aerial drones cooperatively performing work. Optimization of each aerial drone must be done and shared with the other drones likely to be affected in space and time (spatiotemporal). Then the Optimization of other time-dependent subsequent activities may be conducted treating the fleet as a single entity with respect to downstream activities.

These spatiotemporal problems grow geometrically in complexity as the separation in space-time is reduced (i.e. Machine spatiotemporal density increases). Fortunately, with limits in travel speed of each Machine, magnitudes of individual Machine effects diminish in space-time.

P.S. – Thank you for reading this – I am sorry that it has taken this many years and the loss of my dear brother to finally have me write this down. One of Bernie’s favorite expressions in dealing with me was always: “Don’t you write things down?!” He would then always give me his trademark 'toothless' grin and chuckle.