10.06.2016 Views

eldo_user

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Optimization<br />

Optimization of Sweep Simulations<br />

The extracted measure F is a bidimensional object of size defined by the range of the .STEP<br />

statements. Given the values of the variables x (fixed during each simulation), the simulator<br />

returns a value f = f(x; ω (1) , ω (2) ) for each discretized point (ω (1) , ω (2) ) of the domain Ω.<br />

See also:<br />

• “Global and Local Optimization” on page 629<br />

• “Continuous and Discrete Optimization” on page 630<br />

• “Smooth and Non-Smooth Problems” on page 631<br />

• “Multiple-Run Compatibility” on page 632<br />

Inner and Outer Sweep Parameters<br />

Consider the following minimization problem with respect to the x variable, where the design<br />

parameter ω belongs to the interval Ω = [ω start , ω end ].<br />

A formal circuit formulation example is:<br />

.PARAMOPT X=(XINIT, XL, XU)<br />

.STEP PARAM OMEGA OMEGA_START OMEGA_END OMEGA_INCR<br />

.OBJECTIVE<br />

+ EXTRACT_INFO LABEL=f<br />

+ {$MACRO|FUNCTION}<br />

+ GOAL=MINIMIZE<br />

.OPTIMIZE<br />

The strings XINIT, XL, and XU represent: the initial value of the design variable x, its lower<br />

bound, and its upper bound. Refer to “Specifying Design Variables” on page 596 for more<br />

information.<br />

The example optimization statements can be understood in two different ways:<br />

• The optimizer seeks the smallest value of the function:<br />

such that the value of x is within the interval [x l , x u ]. This case is represented in the right<br />

branch of Figure 13-9. Here the design parameter ω is handled as an inner sweep<br />

parameter.<br />

634<br />

Eldo® User's Manual, 15.3

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!