2.4.1 Optimization Target

The result of an optimization will strongly depend on the emphasis that has been layed on individual device categories and their characteristic properties. As an example, a process which produces devices for analog signal processing will lay more emphasis on passive devices like integrated resistors and capacitors.

Figure 2.8 shows the design space where an optimum location has to be found. Each of the dimensions represents a clear design strategy; either Design for Manufacturability (DFM), Design for Performance (DFP), or Design for Reliability (DFR). If we quantify quality in terms of performance, reliability, and manufacturability with measures q_P, q_R, and q_M, respectively, and assign them specific weights (w_P,w_R,w_M), one recognizes that the closer a point is to the origin as described by

$$\frac{1}{\left(q_{tot}\right)^2} = \frac{1}{\left(q_{P}\cdot... ...ight)^2} + \frac{1}{\left(q_{M}\cdot w_{M}\right)^2} \mbox{ ,}$$

the higher is the total quality q_tot of the process which exhibits the corresponding characteristics. The origin represents the ultimate process, since it exhibits optimal properties in terms of DFM, DFP, and DFR. At the same time each point of this space represents a compromise between DFM, DFR, and DFP for the target process.