Examples Overview¶
The framework ships with a set of ready-to-run example use cases located under
userfiles/. Each one demonstrates the distributed optimization workflow on a
different multidisciplinary design problem. The table below summarizes and
compares the available examples along a few criteria.
| Example | Number of Subsystems | Analytical vs. Engineering Problem | Number of Design Variables per Subsystem | Number of Scalar Couplings | ||
|---|---|---|---|---|---|---|
| Local | Shared | Additional (coupling) | ||||
| GeometricProgramming | 3 | Analytical | 3 / 3 / 3 | 0 / 1 / 1 | 2 / 0 / 0 | 3 |
| Sellar | 2 | Analytical | 1 / 0 | 2 / 2 | 1 / 1 | 4 |
| SpeedReducer | 3 | Engineering | 0 / 2 / 2 | 3 / 3 / 3 | 0 / 0 / 0 | 9 |
| SSBJ | 4 | Engineering | 0 / 1 / 3 / 19 | 0 / 0 / 6 / 6 | 5 / 1 / 3 / 1 | 16 |
| TwoBarTruss | 3 | Engineering | 2 / 2 / 2 | 0 / 0 / 0 | 2 / 1 / 2 | 5 |
Reading the table¶
-
Number of design variables per subsystem is split into three categories, each listed in subsystem order (
SS0 / SS1 / ...):- Local — design variables that belong exclusively to a single subsystem (\(x\)).
- Shared — design variables that are shared between neighboring subsystems (\(z\)).
- Additional (coupling) — variables that the decomposition adds as extra design variables to represent the couplings (\(h\)) between subsystems.
The size of a subsystem's full design vector is the sum of its local, shared and additional design variables.
-
Number of scalar couplings counts the total number of scalar coupling relationships (coupling variables plus shared design variables, i.e. the scalar consistency constraints) exchanged between neighboring subsystems.