21.3.6.3. Non-Feasibility of DFSS constraints

In the practical 6-Sigma Design, we don’t guarantee that our design problem has a feasible region. Thus, the convention design tools have frequently failed to satisfy 6-sigma design until the maximum iteration is reached.

Figure 21.84 shows a typical case of non-feasibility of 6-sigma constraints. This represents that the given problem has no feasible optimum fundamentally. Suppose that only one analysis time of RecurDyn is 1hr and the maximum iteration is defined 50 in the problem. Then, it is failed to satisfy 6-sigma constraints. Maybe you check the final design and change the robust index for most violated constraint. Then, you retries to solve it for a long time repeatedly until all constraints are satisfied for the relaxed six-sigma constraints. It is very tedious job.

Figure 21.84 Non-feasibility of DFSS constraints

AutoDesign however automatically checks the feasibility of DFSS constraints in each SAO. Then, the robust indexes are automatically relaxed if it has no feasible region. Figure 21.85 shows the relaxed constraints shown in Figure 21.84.

Figure 21.85 Automatic Relaxation of DFSS constraints