Separation
From DDL Wiki
m (reformatted link) |
Current revision (12:56, 5 May 2009) (view source) (clarified distinction with decomposition) |
||
| Line 1: | Line 1: | ||
==Definition== | ==Definition== | ||
| - | Separation is an [[optimization]] approach that replaces a design optimization problem with a set of subproblems. It then solves each | + | Separation is an [[optimization]] approach that replaces a design optimization problem with a set of subproblems. Thus, it is a type of problem [[decomposition]]. It then solves each |
subproblem once in order to generate a feasible solution. Ideally, separation produces an optimal solution to the original problem. However, not all separations do. | subproblem once in order to generate a feasible solution. Ideally, separation produces an optimal solution to the original problem. However, not all separations do. | ||
| - | + | Separations are similar to decomposition-based design optimization methods, because they all replace a large design optimization | |
| - | problem with a set of subproblems. | + | problem with a set of subproblems. Decomposition-based design optimization methods solve a second-level problem to coordinate the |
| - | subproblem solutions in an iterative manner. | + | subproblem solutions in an iterative manner. Separations do not require this iteration or coordination. |
A separation may use a simple sequence of subproblems, a set of subproblems that are solved in parallel, or a more complex network of subproblems. | A separation may use a simple sequence of subproblems, a set of subproblems that are solved in parallel, or a more complex network of subproblems. | ||
Current revision
Definition
Separation is an optimization approach that replaces a design optimization problem with a set of subproblems. Thus, it is a type of problem decomposition. It then solves each subproblem once in order to generate a feasible solution. Ideally, separation produces an optimal solution to the original problem. However, not all separations do.
Separations are similar to decomposition-based design optimization methods, because they all replace a large design optimization problem with a set of subproblems. Decomposition-based design optimization methods solve a second-level problem to coordinate the subproblem solutions in an iterative manner. Separations do not require this iteration or coordination.
A separation may use a simple sequence of subproblems, a set of subproblems that are solved in parallel, or a more complex network of subproblems.
A separation provides a way to reformulate a design optimization problems and find a good solution. It is important to carefully design the separation to provide a valuable solution, and the quality of the solution that results depends upon the constraints and objectives used in the subproblems. A separation may be able to find a feasible solution in less time.
References
- Karimian, Peyman, and Jeffrey W. Herrmann, "Separating Design Optimization Problems into Decision-Based Design Processes," Journal of Mechanical Design, Volume 131, Issue 1, 011007, January 2009. [1]
