Category: Simulation and Modeling

Can This Just Be My Career?

Kyle

My baseball partner-in-crime Matt sent me this awesome link to a New York Times story I missed about a month ago.

If you want to know how tons of nuclear engineers spend their time, look no further. We run simulations like the one in the article. Except instead of flipping these virtual weighted coins to see how simulated batters fair over the course of their careers, we’re trying to see how simulated neutrons (or x-rays, or electrons, or whatever) fair over the course of their lifetimes (from their creation in the reactor or whatever until their eventual absorption or leakage from the system). Of course, for a binary process (e.g., DiMaggio either gets a hit or he doesn’t) it’s OK to visualize coin-flipping, but I think it’s better to think of the “randomness driver” as the roll of a many, many-sided die–which is why the Men from Mars referred to this clever trick as a Monte Carlo method. (In fact, I used Monte Carlo to write this absurd little simulation, which happens to be about rolling dice.)

Man, writing baseball simulations for a living would maybe be my dream job. Hey, Baseball Prospectus: need any more modelers?

What I Did All Day

Kyle

As it turns out, only this (click to enlarge):

It’s getting to be that time of the semester where scheduling and time management become both more difficult (because the time-uncertainties in end-of-semester-type activities are so much greater) and more important (because of the number and importance of said activities). Case in point: the above plot took me all day to make, and I really only had to generate the dotted line today. It turns out that every time I’ve solved this particular two-region reactor physics problem (at least twice while studying for my qualifying exam), I’ve done it wrong. It took me about two hours to realize my mistake, another two to find and fix it, and another hour to find the additional mistakes I incorporated in moving my solution from one piece of software to another. I can’t tell you how many other items were on my todo list today (including–I’m just realizing–eating lunch), which tasks of course I’m scrambling to do now (well, in about another two minutes, obviously). I’ll try to explain this problem and plot some other time, since they’re kind of interesting and have some bearing on the “What have we got to lose?” modeling question.

Anyway, I just wanted to sort of explain my absence from this space this week and see if anybody had suggestions for dealing with this problem of how you can manage your time when the tasks you’re juggling are both more important and more unpredictable in terms of how much time they take to finish. As far as I can tell, the standard answers include “sleep less” and “forget to each lunch.”

OK, back to work.

…What To Leave Out

Kyle

I got a hat tip from the Freakonomics blog today in reference to a recent post about the Wikipedia article for “real life.” The post accepts the premise that fantasy and reality are complements rather than substitutes (a position Contraria Sunt Complementa naturally supports), and it zeros in on Robin Hanson’s pro-complement point that “fiction can suppress irrelevant detail and emphasize important essences, like a math model” (hyperlink added).

Of course, such suppression is a big part of the game for us researchers who write simulation software. In fact, I’d argue that it’s the game. Not because implementation is trivial (it’s not), but because fantasy has to start from scratch, so to speak; without “suppress[ing] irrelevent detail,” even the fastest model builder never has a chance to catch up. There’s just to much detail to try to capture.

I talk a little bit each week here at CSC about editorial judgment, a subject that–despite going-on six years of engineering training–I have much more real-life experience with than its engineering analog. I’ll make an effort to rectify my recent lack of engineering-judgment coverage in the coming weeks, especially as I try to summon some myself in order to finish up my work on GENIUS by the end of the summer.

I admit to occasionally succumbing to the need for a motivational Post-It note above my workstation. “Simplify, simplify” wouldn’t be a bad motto in light of the addition-by-subtraction nature of modeling. But I’m going to go with the titular reference from the Douglas Adams piece that–almost four years ago now–first got me thinking about the importance of teaching engineers how to decide what to leave out:

“What have we got to lose?”