/sci/ - Science & Math

Goes to 4chan for attention? Makes no sense buddy, but it was a nice that you read what i said. its sad that you didn't add anything to the Comment board.

Ok, Math question.

I currently am learning a bunch of optimization theory stuff and there's one issue I just don't get:

The second-derivative condition for two variables, called the "Second partial derivative test" by wikipedia.

It says, for any stationary point of a function, to be a maximum or minimum, the following condition must be met:

f''11(x,y)*f''22(x,y) - (f''12(x,y))² > 0

If it's < 0 instead, it would be a sattle point. If it equals 0, then it is inconclusive, whatever that means.

What I want to know is why we actually do this. I mean, I can see that we want to test if the two functions multiplied in the first term are either both positive or negative and it makes geometric sense to me that we should have a saddle point if one is positive (thus, the slope is rising) and one negative (thus, the slope is falling), but what if both slopes are positive and we get a number below 0 as the result, what does that mean? How to interpret the second term geometrically?

Please help /sci/.

>>3516207
>weeklystandard
I'm sure that will be a fair and accurate assessment.

Bill Kristol also predicted that the Iraq war would last 6 weeks tops.

>>3220613


you have no clue how big business or for that matter any business runs if you really think that.

It means itself.

>>2078996

And by the same, you mean, not even fucking close.

Idiot.

>>997754
wow, grow up kid, you can't tune a bell or xylaphone

pic related, seriously