STAARS

Abstract

Our goal is to develop methods and tools for designing safe controllers for autonomous, adaptive, and reconfigurable computer systems. To achieve this goal, we propose combining computer science and control theory at the levels of system infrastructure, programming support, and modeling and control techniques. In particular, we explore the relationship between the MAPE-K loop in autonomous computing and continuous and discrete control loops.

Coordinators

Nicolas Marchand (GIPSA-lab)

Eric Rutten (INRIA)

Gwenael Delaval (LIG)

Results

We have identified areas for collaborative research between control theorists and computer scientists. For example, we have proposed a master’s-level course on the integration of probabilistic selection among the solutions of a discrete controller with maximum permissiveness. Other topics include exploring the concept of adaptive discrete control, or identifying control problems in the resource management of multi-tier data centers or high-performance computing systems.

We have engaged the local community on this multidisciplinary topic and organized seminars on control models, programming languages, and adaptive computing infrastructures. Interested researchers in Grenoble are welcome to present their work and exchange ideas. These seminars are held monthly or more frequently. We have organized two workshops on the state of the art in autonomous computing and control, featuring French and European speakers. This project has contributed to the development of several ANR grant proposals. The ANR HPeC project focuses on high-performance embedded computing: we are studying reconfigurable architectures based on FPGAs (Field Programmable Gate Arrays or programmable logic arrays) that are partially dynamically reconfigurable, for autonomous systems, with a case study on drones. In particular, we are working on the control of reconfigurations, using a layered autonomous loop approach.

Notable publications

Eric Rutten, Nicolas Marchand, Daniel Simon. In R. de Lemos; D. Garlan; C. Ghezzi; H. Giese. Feedback Control as a MAPE-K Loop in Autonomic Computing. Software Engineering for Self-Adaptive Systems 3: Assurances, Springer, 2016.

Marin Litoiu, Mary Shaw, Gabriel Tamura, Norha M. Villegas, Hausi Müller, Holger Giese, Eric Rutten, Romain Rouvoy. What Can Control Theory Teach Us About Assurances in Self-Adaptive Software Systems? Software
Engineering for Self-Adaptive Systems 3: Assurances, Springer, 2016.