Can an Actor-Critic Optimization Framework Improve Analog Design Optimization?

arXiv cs.LG / 3/27/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes an actor-critic optimization framework (ACOF) for analog circuit sizing that injects designer-like guidance into simulator-based search rather than treating optimization as a fully black-box problem.
ACOF separates an actor that proposes promising regions of the design space from a critic that evaluates proposals, enforces design legality, and redirects the search when progress stalls.
The approach is designed to remain compatible with standard simulator-based EDA workflows while improving stability and interpretability of the optimization process.
Experiments on multiple test circuits show an average 38.9% improvement in the top-10 figure of merit over the strongest baseline and a 24.7% average reduction in regret, with reported peak gains up to 70.5% FoM and 42.2% lower regret on individual circuits.
Overall, the authors argue that combining iterative “reasoning” (via actor-critic roles) with simulation-driven evaluation yields a more transparent path to automated analog sizing in large, difficult search spaces.

Abstract

Analog design often slows down because even small changes to device sizes or biases require expensive simulation cycles, and high-quality solutions typically occupy only a narrow part of a very large search space. While existing optimizers reduce some of this burden, they largely operate without the kind of judgment designers use when deciding where to search next. This paper presents an actor-critic optimization framework (ACOF) for analog sizing that brings that form of guidance into the loop. Rather than treating optimization as a purely black-box search problem, ACOF separates the roles of proposal and evaluation: an actor suggests promising regions of the design space, while a critic reviews those choices, enforces design legality, and redirects the search when progress is hampered. This structure preserves compatibility with standard simulator-based flows while making the search process more deliberate, stable, and interpretable. Across our test circuits, ACOF improves the top-10 figure of merit by an average of 38.9% over the strongest competing baseline and reduces regret by an average of 24.7%, with peak gains of 70.5% in FoM and 42.2% lower regret on individual circuits. By combining iterative reasoning with simulation-driven search, the framework offers a more transparent path toward automated analog sizing across challenging design spaces.