Language Challenge 2000 -- FAQ

Language Challenge 2000 -- FAQ

[bvoh]

**  Language Challenge 2000  --  Frequently Asked Questions  **


This is a summary of the questions pertaining to the problem
description. To avoid confusion that could critically affect your
solution please examine this list carefully. Your questions have helped
us refine the problem formulation which is available on "Language
Challenge 2000" link on our site.


Q  Is the Standard Atmosphere already in "accepted formulas" or does one
have to do a table look-up, interpolatory approximation, lsq or spline
approximation in order to "guarantee" your targeting error?

A  Limiting the suggestion to using the Standard Atmosphere was
deliberate as we did not want to prejudice design tradeoff
considerations which you have effectively enumerated above. However, not
finding the data wasn't meant to be the eliminating factor, so here are
two relevant sites:

        http://www.digitaldutch.com/atmoscalc
        http://prestwick.simplenet.com/aviation/aviaform.htm

Q  Based on your equations of motion, I would think that you have the
constraint v = sqrt(x'^2 + y'^2) but you don't explicitly say it,
leading the casual observer to think that the velocity along the
trajectory is constant?

A  Indeed, one can never assume it's obvious to everyone. In fact, this
"omission" would have lead to unwittingly missing a target entirely --
problem equations have been amended  accordingly.

Q  Assuming there are two solutions (a high trajectory and a low one) do
you care which one you get?

A  No we don't, but we suspect you might! It is also true that the
initial angle is prescribed and if your optimization happens to stray
into unwanted regions then adjusting algorithm parameters may well be
worth a little extra scrutiny.

Q  It is unclear what exactly "generalized" is supposed to mean.  Is it
supposed to be completely generalized, generalized to problems of this
type, or generalized to this problem's parameters.

A  The second characterization about covers it -- you have to build an
optimization based solution that could be reused by simply substituting
the eom and its performance measure. It was meant to exclude quick trial
& error searches, root solvers or any other technique based on problem
specific apriori knowledge (e.g. solution range, scalar dimension).


Questions regarding the contest should be directed to info@sdynamix.com
with Contest 2000 in the "Subject" line. 

-----------------------------------------------
Modeling * Simulation * Analysis
http://www.sdynamix.com
-----------------------------------------------