Skip to main content
Go to the home page of the European Commission (opens in new window)
English en
CORDIS - EU research results
CORDIS

Synthesis of Reactive Programs

Project description

Towards automatic synthesis of software systems correct by design

Designing systems that work flawlessly from the start could redefine how software is developed. Reactive synthesis refers to an automatic method in hardware design that creates systems capable of ongoing interaction with their environment. However, applying this approach to modern, data-rich systems is challenging owing to the complexity of unbounded data types. Traditional methods cannot efficiently handle these scenarios, failing to meet advanced specifications or manage data reasoning. To address this, the ERC-funded SyReP project plans to develop new techniques to seamlessly integrate logical reasoning with reactive synthesis. Researchers will create symbolic reactive synthesis methods that can reason natively with data. They will also design new constraint-solving and functional synthesis techniques to fit seamlessly into reactive synthesis.

Objective

The automatic synthesis of correct-by-design systems is considered the holy grail of software development.
It promises to revolutionize the traditional process by allowing designers to focus on what the system should
do and not on how to develop an implementation with the desired behavior.
Reactive synthesis – the automatic construction of systems that maintain an ongoing interaction with their
environment – is successfully used in hardware design, as exemplified by case studies like the synthesis of
the AMBA AHB bus controller. However, the algorithmic advances behind this success do not extend to the
complex, cyber-physical, increasingly autonomous systems ubiquitous today. The main obstacle for classical
synthesis methods is that such systems use data domains beyond Boolean variables. Naturally, unbounded
data types render the synthesis problem undecidable. Nevertheless, the importance of reactive program
synthesis, i.e. reactive synthesis in the presence of richer data domains, has sparked a strong interest in
developing techniques for this setting. Current approaches, however, suffer from one or both of two significant
limitations: they are either restricted in the type of correctness specifications they can handle or diverge when
reasoning about unbounded data is required. We aim to address these limitations and profoundly push the
boundaries of reactive program synthesis.
We will do this via a fundamental shift in how reactive synthesis interacts with logical reasoning about
data. Building on recent milestone results by the PI, we will develop symbolic reactive synthesis methods
that reason natively about data. We will develop and employ novel constraint-solving and functional synthesis
methods optimized towards integration in reactive synthesis. Our synthesis techniques will impact many
application domains, such as software for medical devices, mobile applications, and industrial control.

Fields of science (EuroSciVoc)

Programme(s)

Topic(s)

Funding Scheme

HORIZON-ERC - HORIZON ERC Grants

See all projects funded under this funding scheme

Call for proposal

(opens in new window) ERC-2025-COG

See all projects funded under this call

Host institution

CISPA - HELMHOLTZ-ZENTRUM FUR INFORMATIONSSICHERHEIT GGMBH
Net EU contribution
€ 1 996 997,00
Address
STUHLSATZENHAUS 5
66123 SAARBRUCKEN
Germany
Region
Berlin Berlin Berlin
Activity type
Research Organisations
Links
Total cost
€ 1 996 997,00

Beneficiaries (1)