Trust-Region Method with Deep Reinforcement Learning in Analog Design Space Exploration
This work addresses time-to-market challenges in analog circuit design, with incremental improvements in efficiency and industrial deployment.
The paper tackles analog design space exploration by framing it as a constraint satisfaction problem and using a trust-region method with model-based agents, achieving orders of magnitude improvement in search iterations and surpassing human designers on TSMC 5/6nm circuits.
This paper introduces new perspectives on analog design space search. To minimize the time-to-market, this endeavor better cast as constraint satisfaction problem than global optimization defined in prior arts. We incorporate model-based agents, contrasted with model-free learning, to implement a trust-region strategy. As such, simple feed-forward networks can be trained with supervised learning, where the convergence is relatively trivial. Experiment results demonstrate orders of magnitude improvement on search iterations. Additionally, the unprecedented consideration of PVT conditions are accommodated. On circuits with TSMC 5/6nm process, our method achieve performance surpassing human designers. Furthermore, this framework is in production in industrial settings.