Contract-based multi-agent logical controller synthesis under uncertainty

doi:10.1016/j.nahs.2026.101783

Contract-based multi-agent logical controller synthesis under uncertainty

2026-07-07

0

PRE

AI

DOI：10.1016/j.nahs.2026.101783

Original

Original request for help

Save

Abstract

En 中文

We consider the problem of computing distributed logical controllers for multiple interacting system components against a stochastic or adversarial environment via a novel sound and complete contract-based synthesis framework. Based on a discrete abstraction of component interactions as a multi-player game over a finite graph and specifications for all controllable system components given as ω -regular (e.g. LTL) properties over this graph, we co-synthesize contract and controller candidates locally for each controllable component and propose a negotiation mechanism which iteratively refines these candidates until a solution to the given distributed synthesis problem is found. Our framework relies on the recently introduced concept of permissive templates which collect an infinite number of controller candidates in a concise data structure. We utilize the efficient computability, adaptability and compositionality of such templates to obtain an efficient, yet sound and complete negotiation framework for contract-based distributed logical control. We showcase the superior performance of our approach by comparing our prototype tool CoSMo to the state-of-the-art tool on a robot motion planning benchmark suite.

Contract-based multi-agent logical controller synthesis under uncertainty

Abstract

Journal

Organization