Late Breaking Results: Breaking Symmetry- Unconventional Placement of Analog Circuits using Multi-Level Multi-Agent Reinforcement Learning

2025-03-29

AI Summary

Layout-dependent effects (LDEs) degrade analog circuit performance due to process variations. Method: This paper proposes an asymmetric, objective-driven automated placement method that abandons conventional symmetric layout paradigms. It introduces multi-level, multi-agent Q-learning—first applied to analog circuit placement—integrated with accurate LDE modeling and cooperative policy optimization to enable end-to-end learning of placement decisions within an LDE-aware design space. Contribution/Results: Unlike existing rule-based or symmetry-constrained approaches, the framework requires no prior symmetry assumptions, significantly enhancing robustness against process variation. Experimental results at identical technology nodes demonstrate superior performance over state-of-the-art placement techniques—including simulated annealing—in key metrics such as matching accuracy and gain stability, as well as in yield. These findings substantiate the tangible advantages of asymmetric placement for high-performance analog circuit design.

Technology Category

• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Search and Optimization: Learning to Search
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Layout-dependent effects (LDEs) significantly impact analog circuit performance. Traditionally, designers have relied on symmetric placement of circuit components to mitigate variations caused by LDEs. However, due to non-linear nature of these effects, conventional methods often fall short. We propose an objective-driven, multi-level, multi-agent Q-learning framework to explore unconventional design space of analog layout, opening new avenues for optimizing analog circuit performance. Our approach achieves better variation performance than the state-of-the-art layout techniques. Notably, this is the first application of multi-agent RL in analog layout automation. The proposed approach is compared with non-ML approach based on simulated annealing.

Problem

• Mitigate layout-dependent effects on analog circuits
• Explore unconventional analog layout design space
• Optimize analog circuit performance using multi-agent RL

Innovation

• Multi-agent RL for analog layout
• Objective-driven Q-learning framework
• Unconventional asymmetric circuit placement