/sci/ - Science & Math

>>3614784
=( i know

>>3614710
I found the real world stuff "easier" in that it was more interesting, so the work wasn't like an annoyance.

Failure is perfectly fine if you've been learning stuff in it. Demonstrate to them that though things didn't work, you used good reasoning.

>>3614717
while i've done all my work in numerical methods, i can't say i actually own any books strictly on the subject.

>>3614763
an unfortunate distraction.

>>3614646
How do you mean "no exact method"? Is the error strictly due to computational intractability of finding a nonlinear global optimization for a large number of DOFs?

Controls ignorance showing here, but hopefully you get what I'm asking. It sounds similar to adjoint methods.

>>3614511
Well, an __-infinity norm is just the maximum value in some space... minimizing the H-infinity is just minimizing the maximum error for a control law. AFAIK, this is really just an optimization problem for a large number of variables. I'd have to defer to my ME buddy over there for more details.

>>3614426
Totally. I love controls. Optimal controls has been the most fun topic for me to study outside of my own area.

>>3614395
fair enough.

>>3614378
>Is there anything else that you considered when looking at grad schools? There must have been a time when you weren't sure about what you wanted to research.

There are some people that look at rankings and such. I think that's a bad idea.

There was a time when I didn't know I wanted to do research. Then I started to learn about the field I'm currently in circa my junior year, and realized I wanted to work in that, and that it required grad education.

Just realize that if you're serious about grad school, it can entail many many hours for potentially low stipend pay, high stress, and low thanks. After that, you might find you don't like it at all. I mean, an MS is pretty easy, so go for it.

>>3614318
... did you mean pessimistic?

Sure, some have plans. But I've seen tons of grad students show up with literally no idea what area they want to do research in. These people signed on for a 5 year program in research without any desire to do anything in particular. They inevitably do incredibly mediocre work.

So yeah, some people have plans. But "I want to go to grad school but I don't know what area" is not a plan. It just reeks of "my parents went to grad school and I want to please them."

>>3614273
>I wouldn't consider my science education complete without at least a masters in an EE or Physics related field.

What are you basing that on? That it "sounds cool"? You literally don't even know what it is you want to know. You have to figure that out before you can know if you want more education in it.

>>3614280
I wouldn't necessarily say that. I mean, to be pragmatic, you probably shouldn't go into some unemployable area like philosophy. But if you're a smart guy and enjoy math and physics, then why not go to university and get exposed to things you might really like and be able to make a career of it?

I'd say the opposite of your statement is true... what 17 year old can intelligently say what they'd like to do for a scientific career?

>>3614217
I've studied it, but I have no real interest in low Re fluids. Extremely boring for a computational person.

>OP where do you get your news from? Which sites/magazines/journals do you subscribe to?
Nothing out of the ordinary. yahoo news, slashdot, scientific american for mainstream stuff. Otherwise, AIAA newsletters and regular scientific journals in my field.

>>3614155
>Am I supposed to know research scientists as a 3rd year student? I haven't even delved deeply into my subject.
Then you have no reason to be considering grad school.

>>3614161
>How much asskissing/politics is involved in your line of 'work'?
None. This isn't liberal arts. You're either a capable student/researcher or you're not.

Of course, there's still room to have good social skills, as in any workplace.

>>3614111
=/ Um, what?

>>3614139
Fluid mechanics is generally pretty challenging, yeah.

>>3614088
Yes, if and only if you directly contact professors you're interested in, and convince them to take you on.

>>3614037
Directly contact a professor that you want to work for, and convince them that you will do a good job and you're passionate about their work.

If you have a low GPA because you're stupid or lazy or uninterested in the field, then forget about it.

>>3614055
Perfect answer.

>>3614025
Look at the website for every department you'd even remotely consider, and look at the research interests / CVs of the professors. You should already know what you want to research. Then, contact them.

>>3613978
To apply to *any* engineering discipline... really means a wide range of applied mathematics.

Differential equations, real and complex analysis, statistics, linear algebra, numerical methods, perturbation theory, variational calculus.

>>3613786
I think Star Trek was a major influence for me, so maybe not that retarded. I ended up in a field I had no idea existed. Studying aerospace exposes you to an awful lot of applied math and science concepts. So sure, maybe you won't make it in asteroid mining, but you'll find something.

>>3613262
LOL. Any decent school tends to fully fund their engineering PhD students. Kind of a non-issue.

>>3613236
no idea.

>>3613171
Possibly extremely easy for EE undergrads. Probably cover the basics of finite difference methods, ODE integrators (runge kutta etc), numerical linear algebra (solving systems of equations and such)

>>3613165
Also, the CUDA SDK has a hard-sphere collision code in it that works quite well. That'd be a great place to start to move to SPH on GPUs.

>>3613127
Well, you can get away with doing it only as SPH if you want, but FEM can handle it just as well. SPH is just a lot easier for deforming grids (as there is no grid.) I don't have any texts on that matter, but the original SPH papers by J J Monaghan are probably the best place to start.

>>3613100
... what did you expect? No pity from me for going into work/research you didn't know anything about.

>>3613101
Yeah, don't try to cram or memorize shit. Learn to derive everything from first principles. Also, since physics, find a job?

>>3613086
Well, obviously SpaceX has accomplished the most. I guess Blue Origin really interests me out of their sense of mystery, and I eagerly wait for them to announce they did something really awesome.