How to relate AWR v16 optimizers to new AWR v22.1 optimizers

In the AWR Design Environment platform V22.1 software, the optimizer list is reorganized and reduced to simplify selection. Following optimizer table maps V16 software optimizers by names to their equivalent in V22.1 software:

# A new Pointer: Hybrid optimizer method is introduced, which combines the best performance of several optimizers to produce a top-performing algorithm on a very wide range of problems.

Note that all optimizers except Lineup and Gradient:

• Can run in parallel
• Can use discrete variables and step sizes
• Now require variable constraints

The new optimizers are tested and found to perform as good as or better than the older optimizers they are mapped to, with no loss of capability on a wide range of problems. Note, however, that there may be exceptions. If an optimizer that is no longer available is required to achieve the best performance on a particular problem, contact Cadence Technical Support for assistance.

Unconstrained variables defined in the older projects must be updated with constraints to work with the new optimizers. Following are the steps to quickly add constraints:

1. Open Variable Browser (choose View > Variable Browser).
2. Type "1" in the Optimize column header to filter on optimized variables.
3. Shift+select all displayed variables.
4. Select the Constrained checkbox for the last selected variable to update the checkbox for all selected variables.
5. Enter lower and upper bounds for the variables as a percentage or absolute delta of the current value by following the lower and upper bound value with a "%" or "#" sign, respectively. For example, to enter a lower bound that is 10 percent lower than the current value for all the variables, type "10%" in the Lower column for the last selected variable.

Now that you have read the article, answer the following questions in the comment box:

Q1. What is the primary objective of optimization techniques?

a) To find the best solution

b) To find any solution

c) To find a quick solution

d) To find the most complex solution

Q2. Which new optimization method in AWR v22.1 combines the strengths of multiple optimizers to produce high performance across a wide range of problems?

a) Pointer – Hybrid

b) Robust Simplex

c) Lineup Optimizer

d) Gradient Optimization

Team SimTech

Cadence Design Systems