Skip to content

Align aircrafta with SIF and add aircrafta2 - #479

Open
timholy wants to merge 2 commits into
JuliaSmoothOptimizers:mainfrom
timholy:teh/aircrfta2
Open

timholy wants to merge 2 commits into
JuliaSmoothOptimizers:mainfrom
timholy:teh/aircrfta2

Conversation

@timholy

@timholy timholy commented Oct 20, 2026

Copy link
Copy Markdown
Contributor

Gives aircrafta the same starting point as the SIF version.

Adds aircrafta2, which SIF calls the corrected version of aircrafta. This problem differs both in coefficients and the form of the product.

AIRCRFTA.SIF (Ph. Toint's SIF input of problem 9 in J. J. Moré, "A
collection of nonlinear model problems") fixes the three controls
through their bounds, elevator = 0.1, aileron = 0 and rudderdf = 0,
and starts each of them at its fixed value, with the five states at
zero. The ADNLPProblems and PureJuMP implementations started the
elevator at 0, outside its bounds 0.1 <= elevator <= 0.1.

The starting point is now (0, 0, 0, 0, 0, 0.1, 0, 0) in both
implementations. It satisfies the variable bounds, and the constraint
values there are (0, -2.837, 0, -0.1168, 0).

See JuliaSmoothOptimizers#37.

Assisted-by: Claude Opus 5.5 <noreply@anthropic.com>
aircrfta2 is the aircraft stability problem of AIRCRFTA2.SIF. It
differs from aircrfta (AIRCRFTA.SIF) in two places: the elevator
coefficient in the fourth equation is -0.168 rather than -1.168, and
the bilinear terms with coefficients 63.5 (first equation) and 1.132
(third equation) are products of the aileron and the angle of attack
rather than of the pitch rate and the angle of attack.

With the controls fixed at their SIF values, the five equations have a
solution, so the best known upper bound on the constant objective is 0.

See JuliaSmoothOptimizers#37.

Assisted-by: Claude Opus 5.5 <noreply@anthropic.com>

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant