# Simulated Education: STEM must change!

-October 24, 2014

The linear low-hanging fruit has all been picked and it's time for STEM (science technology engineering and mathematics) education to leap into the 21st century of complex systems. Why? Because STEM education focuses on archaic problem solving techniques.

The tools used to approach the complex, nonlinear problems faced in modern society don't fit the educational techniques that have been used for the last few hundred years. Sure, computing devices play a role in the classroom, but test techniques and homework assignments, even projects, no longer resemble the work of practicing professionals.

The problems that will confront scientists and engineers who begin their undergraduate educations this year will be complex. Not just complicated, but nonlinear, complex problems whose solutions can't be derived through linear approximation.

Consider a typical homework problem:

In 1960, Mrs. Maxwell Rogers of Tampa, Florida, reportedly raised one end of a car that had fallen onto her son when a jack failed. If her panic-lift effectively raised 900 lb. of the 3600. lb car by 2 in., how much work did she do? (from Fundamentals of Physics, 5th edition, by D. Halliday, R. Resnick, and J. Walker, Wiley).

This is a simple, plug-and-chug problem solved by assembling the appropriate mathematical physics on paper with a pencil. Each such problem results in an opportunity for students to refine not just their understanding of the science, but their ability to practice it by providing the answer.

A typical classroom of the 20th century.

The problem with this approach to science-and-engineering education is that the behavior of the vast majority of physical systems can't be resolved into a tidy equation that can be written on a sheet of paper. Back in the 20th century, teachers could guide students through solvable problems and then, if there was time at the end of the term, provide ways to approximate solutions to problems whose solutions can't be written down in a few equations. Judging by the science and technology developed in the 1900s, the system worked really well.

Linear approximations worked great for Morse code on long cables, but they're so useful for DP-QPSK (dual polarization quadrature phase shift keyed) signals on optical fibers. All those clever analog approaches to solving electromechanical problems demonstrated unqualified genius, but now we program nonlinear IIR (infinite impulse response) filters on microprocessors that perform dozens of operations in a nanosecond; we use MEMS (micro-electromechanical systems) to light up pixels on movie screens and to switch signals in optical and quantum computers.