Real-Time Scheduling of Reconfigurable Distributed Embedded Systems with Energy Harvesting Prediction

Housseyni, Wiem; Mosbahi, Olfa; Khalgui, Mohamed; Chetto, Maryline

doi:10.1109/ds-rt.2016.30

Cited by 6 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Under that condition, energy-greedy algorithms exploit the available slack in the system by procrastinating tasks and charging the battery as much as possible. In contrast, the computation-greedy algorithms give priority to execute the pending workload, and charge the battery only for the shortest time, which guarantees the execution of the next computational unit (Moser et al, 2006;Bambagini, 2014;Chandarli, 2014;Housseyni et al, 2016).…”

Section: Real Time Schedulingmentioning

confidence: 99%

AI-Based Methods to Resolve Real-Time Scheduling for Embedded Systems

Boutekkouk

2022

International Journal of Cognitive Informatics and Natural Intelligence

View full text Add to dashboard Cite

Artificial Intelligence is becoming more attractive to resolve nontrivial problems including the well known real time scheduling (RTS) problem for Embedded Systems (ES). The latter is considered as a hard multi-objective optimization problem because it must optimize at the same time three key conflictual objectives that are tasks deadlines guarantee, energy consumption reduction and reliability enhancement. In this paper, we firstly present the necessary background to well understand the problematic of RTS in the context of ES, then we present our enriched taxonomies for real time, energy and faults tolerance aware scheduling algorithms for ES. After that, we survey the most pertinent existing works of literature targeting the application of AI methods to resolve the RTS problem for ES notably Constraint Programming, Game theory, Machine learning, Fuzzy logic, Artificial Immune Systems, Cellular Automata, Evolutionary algorithms, Multi-agent Systems and Swarm Intelligence. We end this survey by a discussion putting the light on the main challenges and the future directions.

show abstract

Section: Real Time Schedulingmentioning

confidence: 99%

AI-Based Methods to Resolve Real-Time Scheduling for Embedded Systems

Boutekkouk

2022

International Journal of Cognitive Informatics and Natural Intelligence

View full text Add to dashboard Cite

show abstract

“…Our solution's runtime is typical of heuristics, i.e., tens of milliseconds (see Section V), whereas meta-heuristics normally require seconds or minutes to produce a solution [5]. In list scheduling [6]- [8], individual tasks are sorted in a priority list and assigned, in sequence, to the earliest available unit that fits their resource request. Priorities can be assigned statically or dynamically based on di erent characteristics, e.g., execution time, resource occupancy.…”

Section: Rmentioning

confidence: 99%

Efficient Scheduling of FPGAs for Cloud Data Center Infrastructures

Bertolino

Pacalet

Apvrille

et al. 2020

2020 23rd Euromicro Conference on Digital System Design (DSD)

View full text Add to dashboard Cite

In modern cloud data centers, reconfigurable devices can be directly connected to a data center's network. This configuration enables FPGAs to be rented for acceleration of dataintensive workloads. In this context, novel scheduling solutions are needed to maximize the utilization (profitability) of FPGAs, e.g., reduce latency and resource fragmentation. Algorithms that schedule groups of tasks (clusters, packs), rather than individual tasks (list scheduling), well match the functioning of FPGAs. Here, groups of tasks that execute together are interposed by hardware reconfigurations. In this paper, we propose a heuristic based on a novel method for grouping tasks. These are gathered around a high-latency task that hides the latency of remaining tasks within the same group. We evaluated our solution on a benchmark of almost 30000 random workloads, synthesized from realistic designs (i.e., topology, resource occupancy). For this testbench, on average, our heuristic produces optimum makespan solutions in 71.3% of the cases. It produces solutions for moderately constrained systems (i.e., the deadline falls within 10% of the optimum makespan) in 88.1% of the cases.

show abstract

“…The research reported in [112] extended to [35] deals with the migration of reconfigurable real-time tasks under harvesting energy constraints. The work reported in [169] predicts the harvesting energy availability for improving the performance of real-time scheduling of reconfigurable tasks in multiprocessor architecture. The research reported in [170] deals with the real-time scheduling of reconfigurable sporadic tasks in a multiprocessor architecture.…”

Section: B Multiprocessor Scheduling Of Reconfigurable Tasks Under Ementioning

confidence: 99%

On Reconfiguration Theory of Discrete-Event Systems: From Initial Specification Until Final Deployment

Khalgui

Mosbahi

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

This paper presents an overview on different research activities that we did in the recent decade for developing reconfigurable discrete-event systems (RDESs) from initial high-level specification according to user requirements until final low-level deployment in target hardware components. A reconfiguration is any run-time scenario that adapts the system's behavior to any evolution in the related environment by adding or removing services or also configuring their parameters, i.e., their frequency, execution time, or also their location in the considered hardware architecture. Since the development of distributed RDESs under functional and extra-functional constraints is required by experts, we propose a complete methodology that deals first with their initial design with a new general profile named R-UML extending unified modeling language (UML) or also with new specific technology-oriented profiles, the validation of the related models with a new language R-OCL extending object constraint language (OCL), before their transformation to formal formalisms, such as Petri nets, timed automata, or B method for simulation or also formal verification of different properties. The checked models are transformed into OS reconfigurable tasks in the operational level, before applying a co-design methodology under functional, real-time, memory, and energy constraints for minimizing redundancies in tasks and for optimizing the composition of software and hardware parts together. We describe technology-oriented solutions for the scheduling of distributed RDESs by parameterizing tasks and their exchanged messages between multi-speed networked processors. We finish with the real low-level deployment on target hardware devices before applying useful software and hardware tests for checking the delivered system quality. These contributions, published in different journal and conference papers, are applied to different applications in medicine, wireless sensor networks, transportation systems, manufacturing industry, smart grids and microgrids, embedded technologies, and so on. We find significant gains in terms of the system reactivity and flexibility under related constraints. INDEX TERMS Discrete-event system, modeling and verification, design and validation, implementation and scheduling, real-time and energy aware scheduling, test and simulation, co-design and deployment, application. I. INTRODUCTION Automatic flexibility becomes a required criterion for dynamic system adaptability to any possible evolution in related environment. A reconfigurable system is considered in this paper as a discrete-event system (RDES) that adapts its behavior to recover faults or also improve required performance at run-time. It is a set of wired/wireless networked devices that run independent/dependent control software tasks exchanging messages on the network [1], [2]. These tasks and consequently the related messages should be executed under functional, memory, real-time, and energy constraints described in user requirements [3]-[5]. A reconfi...

show abstract

Real-Time Scheduling of Reconfigurable Distributed Embedded Systems with Energy Harvesting Prediction

Cited by 6 publications

References 16 publications

AI-Based Methods to Resolve Real-Time Scheduling for Embedded Systems

AI-Based Methods to Resolve Real-Time Scheduling for Embedded Systems

Efficient Scheduling of FPGAs for Cloud Data Center Infrastructures

On Reconfiguration Theory of Discrete-Event Systems: From Initial Specification Until Final Deployment

Contact Info

Product

Resources

About