The Sellar problem is an analytical multi-disciplinary design problem. The objective function f is to be minimized with
 respect to the variables z1, z2 and x1 while being subjected to g1, g2 and the disciplines y1 and y2.

D1: y1 = z1^2 + x1 + z2 -0.2*y2
D2: y2 = y1^0.5 + z1 + z2
F1: f = x1^2 + z^2 + y1 + e^(-y2)
G1: g1 = y1/3.16 - 1
G2: g2 = 1 - y2/24

More information on the Sellar problem can be found in literature, e.g. on:
https://arc.aiaa.org/doi/abs/10.2514/6.1996-714, 
https://www.researchgate.net/publication/2759746_Response_Surface_Based_Concurrent_Subspace_Optimization_For_Multidisciplinary_System_Design or
http://openmdao.org/releases/0.2.5/docs/mdao/intro.html

The files in the subdirectories represent and illustrate Repository Connectivity Graphs (RCGs), Fundamental Problem
Graphs (FPGs) and MDO Data/Process Graphs (MDG/MPG). These graphs were developed for a new software system called
KADMOS.

In the RCG files additional tools are present, such as A, B, C. Some of these tools are removed, other are introduced to include
functions with different roles in the repository. However, none of these tools change the actual problem to be solved.

Two implementations of the Sellar case are possible, "math" or "competences". The math configuration creates mathematical relations
for all the Sellar functions, while the competences configuration assumed the Sellar functions to be implemented as design
competences (black boxes).