On self-adaptive resource allocation through reinforcement learning

Panerati, Jacopo; Sironi, Filippo; Carminati, Matteo; Maggio, Martina; Beltrame, Giovanni; Gmytrasiewicz, Piotr J.; Sciuto, D.; Santambrogio, Marco D.

doi:10.1109/ahs.2013.6604222

Cited by 14 publications

(11 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, the authors of [18] present an approach to support the adaptation process of CPS based on run-time generation of verified system configurations, while in [19] the temporal costs of an autonomic manager that performs on-line verification for a specific application are analysed. Moreover, [20] presents an architecture of a middleware that supports time-deterministic reconfiguration in distributed soft real-time environments, while [21] evaluates reinforcement learning adaptation policies through a set of experiments. All the aforementioned works present general approaches that have been applied in a specific application scenario.…”

Section: Proofmentioning

confidence: 99%

Managing safety and mission completion via collective run-time adaptation

Bozhinoski

Garlan

Malavolta

2019

Journal of Systems Architecture

View full text Add to dashboard Cite

Mobile Multi-Robot Systems (MMRSs) are an emerging class of systems that are composed of a team of robots, various devices (like movable cameras, sensors) which collaborate with each other to accomplish defined missions. Moreover, these systems must operate in dynamic and potentially uncontrollable and unknown environments that might compromise the safety of the system and the completion of the defined mission. A model of the environment describing, e.g., obstacles, no-fly zones, wind and weather conditions might be available, however, the assumption that such a model is both correct and complete is often wrong. In this paper, we describe an approach that supports execution of missions at run time. It addresses collective adaptation problems in a decentralized fashion, and enables the addition of new entities in the system at any time. Moreover, it is based on two adaptation resolution methods: one for (potentially partial) resolution of mission-related issues and one for full resolution of safety-related issues.

show abstract

Section: Proofmentioning

confidence: 99%

Managing safety and mission completion via collective run-time adaptation

Bozhinoski

Garlan

Malavolta

2019

Journal of Systems Architecture

View full text Add to dashboard Cite

show abstract

“…Panerati et al [16] adopted the reinforcement learning-based approach to enable self-adaptive resource allocation. Although the core technique is similar, we consider an additional factor which is the complexity of the learning approach whenever multiple multiple adaptation managers are involved.…”

Section: Related Workmentioning

confidence: 99%

Decentralized Planning for Self-Adaptation in Multi-cloud Environment

Ismail

Cardellini

2015

Communications in Computer and Information Science

View full text Add to dashboard Cite

Abstract. The runtime management of Internet of Things (IoT) oriented applications deployed in multi-clouds is a complex issue due to the highly heterogeneous and dynamic execution environment. To effectively cope with such an environment, the cross-layer and multi-cloud effects should be taken into account and a decentralized self-adaptation is a promising solution to maintain and evolve the applications for quality assurance. An important issue to be tackled towards realizing this solution is the uncertainty effect of the adaptation, which may cause negative impact to the other layers or even clouds. In this paper, we tackle such an issue from the planning perspective, since an inappropriate planning strategy can fail the adaptation outcome. Therefore, we present an architectural model for decentralized self-adaptation to support the crosslayer and multi-cloud environment. We also propose a planning model and method to enable the decentralized decision making. The planning is formulated as a Reinforcement Learning problem and solved using the Q-learning algorithm. Through simulation experiments, we conduct a study to assess the effectiveness and sensitivity of the proposed planning approach. The results show that our approach can potentially reduce the negative impact on the cross-layer and multi-cloud environment.

show abstract

“…Adaptation policies, which can be implemented in kernel [5,25] and user-space [19], can exploit the availability of performance measurements and QoS requirements to affect applications' execution. HRM is tightly integrated with the Linux kernel since its primary goal was to export application-specific performance measurements to the kernel-space.…”

Section: Performance Monitoringmentioning

confidence: 99%

“…Moreover, these works exploited different decision-making techniques spanning from heuristics [5,9,21,25] to machine learning [7,19], control theory [22], a mix of them [11].…”

Section: Related Workmentioning

confidence: 99%

A performance-aware quality of service-driven scheduler for multicore processors

2014

Self Cite

View full text Add to dashboard Cite

In the latest decade, the IT industry shifted from single to multicore processors. Multicore processors require better support from operating systems and runtimes to allow applications to achieve predictable performance and guarantee quality of service (QoS). Finding a proper schedule to yield the specified performance for single and multi-threaded applications can be cumbersome; dealing with multi-programmed workloads may be even worse.We present a performance-aware QoS-driven scheduler for multicore processors, which exploits the availability of runtime application-specific performance measurements to determine a suitable allotment of cores for multi-programmed workloads so as to achieve the desired level of QoS. The proposed scheduler is meant to be implemented in user-mode and harnesses an auto-regressive moving average performance model to put in a relationship performance measurements and resource allocation and is capable of embodying applications' characteristics such as execution phases.

show abstract

On self-adaptive resource allocation through reinforcement learning

Cited by 14 publications

References 14 publications

Managing safety and mission completion via collective run-time adaptation

Managing safety and mission completion via collective run-time adaptation

Decentralized Planning for Self-Adaptation in Multi-cloud Environment

A performance-aware quality of service-driven scheduler for multicore processors

Contact Info

Product

Resources

About