# Compositional Synthesis via a Convex Parameterization of Assume-Guarantee Contracts

arXiv: https://arxiv.org/abs/2002.06494 · HTML: https://mlsanity.com/paper/2002.06494/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Feb 2020.

## Built on by (most cited first, 2 of 2)

- [Safe Reinforcement Learning using Action Projection: Safeguard the Policy or the Environment?](https://mlsanity.com/paper/2509.12833/) (2025)
- [Formal Control for Uncertain Systems via Contract-Based Probabilistic Surrogates (Extended Version)](https://mlsanity.com/paper/2506.16971/) (2025)

## Abstract (arXiv)

We develop an assume-guarantee framework for control of large scale linear (time-varying) systems from finite-time reach and avoid or infinite-time invariance specifications. The contracts describe the admissible set of states and controls for individual subsystems. A set of contracts compose correctly if mutual assumptions and guarantees match in a way that we formalize. We propose a rich parameterization of contracts such that the set of parameters that compose correctly is convex. Moreover, we design a potential function of parameters that describes the distance of contracts from a correct composition. Thus, the verification and synthesis for the aggregate system are broken to solving small convex programs for individual subsystems, where correctness is ultimately achieved in a compositional way. Illustrative examples demonstrate the scalability of our method.