Conclusion

This chapter first defines which are the targets of optimization, and then it applies the mathematical theory of chapter 4 (page ) to the optimization of real circuits.
It has been shown how the only mono-objective optimization feasible by means of transistor dimensions trimming is the delay minimization, since both the minimization of area and power consumption lead the quasi-obvious solution of all transistor at the minimum width allowed by the technology or by the designer.

Regarding the multi-objective optimization a method that permits to tackle several optimization policies has been presented. This method permits to take into account all the variables, even whether they are incommensurable among themselves; by means of a normalization all the targets to be optimized can be combined in a single objective function, with a relative ``agreement'' level.
Moreover with this way of combining the several targets into one objective function, the introduction of constraints is as simple as it is in a mono-objective optimization.