21.3.6.4. Random Variables

As one uses Robust Design Optimization, he can define the design parameters from variable and constant in the window of Design Variable. However, it is somewhat complicated than Design Optimization.

First, design parameters can be Variable or Constant in the similar way of design optimization. Second, the defined Variables or Constants can be Deterministic or Random values. Although the design problems can be robust design or DFSS, their design variables may be deterministic or random. The random variables have variations but the deterministic does not. When he uses random variables, he should define the variations from STD or COV types. STD represents standard deviation type. So, user defines the value of variations directly. COV represents coefficient of variation type. Now, let’s see the difference between two variation types.

If DV(1) is a random variable and its’ variation is DV(2), then its’ variation is internally defined as \(\delta \left( 1 \right)=\pm 0.01\) during optimization process. However, if DV(2) is a random variable and its’ variation is COV=0.01, then its’ variation is defined as \(\delta \left( 2 \right)=\pm 0.01\left| DV\left( 2 \right) \right|\) in the optimization process. It is noted that \(\delta \left( 2 \right)\) is changed with DV(2) in the design process.