resource-aware scheduling

To address the task offloading scheduling challenge in edge-cloud computing under resource constraints, this paper systematically surveys cooperative scheduling techniques that jointly optimize low latency, high quality of experience (QoE), and strong fault tolerance. We propose a fine-grained taxonomy based on algorithmic attributes—such as heuristic and metaheuristic approaches—and establish a cross-layer unified evaluation framework to enable precise alignment between scheduling strategies and application scenarios. By integrating QoS modeling, fault-tolerance mechanisms, and multidimensional comparative analysis, we synthesize a comprehensive, structurally coherent knowledge graph for edge-cloud task scheduling. Our key contributions include: (i) a novel classification paradigm grounded in algorithmic characteristics; (ii) a scalable, scenario-aware evaluation framework; and (iii) a systematic identification of open challenges and future research directions—thereby providing both theoretical foundations and practical guidance for designing adaptive scheduling algorithms.

This paper addresses the task scheduling challenge in geo-distributed computing, arising from network heterogeneity, heterogeneous resource pricing, and imbalanced computational capacity. It systematically surveys scheduling techniques across four paradigms: cloud, edge, cloud–edge collaboration, and high-performance computing (HPC). The study introduces the first unified taxonomy covering all four environments, grounded in three core objectives—performance, fairness, and fault tolerance—and identifies cross-cutting challenges including cross-domain latency-sensitive scheduling, multi-regional cost optimization, and elastic fault tolerance. Through bibliometric analysis and qualitative comparative evaluation, it classifies and assesses state-of-the-art approaches—including multi-objective optimization, game-theoretic models, reinforcement learning, and heuristic algorithms. The work traces the technical evolution of scheduling research and proposes six future directions: AI-native schedulers, carbon-aware scheduling, among others—thereby providing theoretical foundations and practical guidance for building adaptive, sustainable distributed scheduling systems.

Existing deep learning training schedulers for heterogeneous AI accelerator clusters (e.g., GPU/TPU) lack fine-grained modeling of hardware-level performance heterogeneity, leading to suboptimal resource utilization and scheduling inefficiency. Method: We propose Hadar, a task-level scheduler, and its enhanced variant HadarE. Their core innovation lies in explicitly incorporating task-level performance heterogeneity into the optimization framework. HadarE further introduces a multi-replica, cross-node concurrent training mechanism to enable coordinated execution across heterogeneous devices while preserving inference accuracy. The approach integrates primal-dual online scheduling, heterogeneous performance feature modeling, and trace-driven evaluation. Results: Experiments show that Hadar achieves 1.20× speedup over the state-of-the-art scheduler Gavel. HadarE reduces total training time by 50% on AWS and 80% on an on-premises cluster, improves resource utilization by 1.45×, and maintains higher model accuracy.

This paper addresses the QoS-aware scheduling problem for periodic Directed Acyclic Graph (DAG) tasks under strict end-to-end latency constraints in heterogeneous, partially occupied Mobile Edge Computing (MEC) environments. We propose a hyperperiod-feasible scheduling method tailored for pre-emptive MEC systems. Our approach innovatively extends the HEFT algorithm to dynamically identify processor idle intervals, while jointly modeling task periodicity, precedence constraints, and resource limitations. We further introduce VM-level task binding and residual computing capacity awareness to achieve fine-grained task-resource matching without compromising the stability of existing workloads. Experimental results demonstrate that the proposed method significantly improves schedulability, load balancing, and resource utilization. It effectively supports high-density periodic DAG workloads while guaranteeing hard real-time QoS requirements.

This work addresses workflow scheduling for cyber-physical systems (CPS) in the edge–cloud continuum, where heterogeneous edge devices face coupled constraints on computation, communication, energy, and sensing/actuation capabilities. The objective is to minimize end-to-end latency under these multi-dimensional resource limitations. To this end, we formulate a Mixed-Integer Linear Programming (MILP) model that explicitly captures the heterogeneity of compute, network, energy, and CPS-specific resources. We further design an enhanced heuristic algorithm to efficiently solve the NP-hard scheduling problem. Compared to state-of-the-art approaches that neglect realistic multi-constraint interactions, our method achieves an average latency reduction of 13.54% on real-world use cases and up to 33.03% on synthetic workflows. The solution demonstrates both significant performance gains and strong scalability. This work provides a theoretically grounded, empirically validated framework and a practical methodology for intelligent workflow orchestration in resource-constrained edge–CPS environments.

This work addresses the scheduling of partially replicable task chains (e.g., SDR communication standards) on heterogeneous multicore platforms, jointly optimizing throughput and power consumption. We formulate the problem—uniquely integrating partial replicability and big-little core co-scheduling—as a dual-resource pipelined workflow scheduling problem. To solve it, we propose: (i) FERTAC/2CATAC, a near-optimal greedy algorithm; and (ii) HeRAD, an optimal dynamic programming algorithm—both unifying pipelined and replication-based parallelism. Experiments show that FERTAC/2CATAC achieves average cycle times within <10% of HeRAD’s, with at most two additional cores overhead. On the StreamPU platform and in real-world DVB-S2 deployments, our approach attains >92% of theoretical peak throughput, significantly improving energy efficiency and scalability.

This paper addresses energy-efficient task offloading for deadline-constrained IoT applications in multi-access edge computing (MEC), aiming to maximize total energy savings across end devices through joint offloading decision-making and communication-computation resource allocation. Method: We propose a general offloading model supporting computation offloading to non-co-located edge servers and formulate the problem as an integer nonlinear program (INLP). To solve it efficiently, we design the first approximation algorithm—Graph-based Matching Algorithm (GMA)—based on tripartite graph matching, supported by linear programming relaxation, graph-theoretic modeling, and LP rounding. Contribution/Results: GMA is theoretically proven to achieve a $(1-alpha)/(2+varepsilon)$ approximation ratio. Experiments demonstrate that GMA attains, on average, 97% of the optimal energy savings, significantly outperforming baseline methods. It constitutes the first provably efficient, theoretically guaranteed energy-saving solution for latency-sensitive MEC offloading.

In the post-Dennard era, embedded systems face intricate trade-offs between energy efficiency and latency, rendering traditional heuristic methods ineffective in navigating the high-dimensional, non-smooth scheduling space. This work proposes a Gaussian process-based multi-objective Bayesian optimization framework to automatically discover Pareto-optimal scheduling strategies that balance energy consumption and execution time on heterogeneous multicore architectures. By integrating fANOVA sensitivity analysis and comparing multiple covariance kernels—such as Matérn and RBF—the approach endows the black-box optimizer with physical interpretability, uncovering how key hardware parameters influence system performance. Experimental results demonstrate that the method efficiently approximates the Pareto front, significantly advancing both the automation of scheduling and the understanding of underlying hardware behaviors.

This study addresses the problem of scheduling precedence-constrained workflows in heterogeneous data centers with the dual objectives of meeting strict deadlines and minimizing carbon emissions. The authors formally prove that this problem is NP-hard and does not admit a constant-factor approximation algorithm. To tackle it, they propose a novel algorithm, Carbon-aware Workflow Mapping (CWM), which integrates dynamic programming with heuristic strategies to jointly optimize task mapping and scheduling. CWM explicitly models node-level energy heterogeneity and temporal fluctuations in renewable energy availability. Experimental results demonstrate that under a deadline constraint set at twice the baseline makespan, CWM reduces median carbon cost by 42% compared to the state-of-the-art CaWoSched algorithm, substantially improving carbon efficiency.

This study addresses the trade-off between energy efficiency and latency for AI/ML workloads in multi-instance GPU (MIG) environments by proposing a dynamic repartitioning scheduling framework tailored to individual MIG instances. The framework integrates scheduler selection with a reinforcement learning–based dynamic repartitioning strategy, marking the first application of reinforcement learning to MIG reconfiguration decisions. It uncovers optimal GPU partitioning patterns under varying temporal and queue-state conditions, enabling predictive and automatic adjustments. Experimental evaluation using real-world diurnal workload traces from data centers demonstrates that the proposed approach improves the combined metric of energy consumption and task latency by 26%, 31%, and 68% compared to twice-daily repartitioning, static partitioning, and no partitioning schemes, respectively.

This work addresses the challenge of resource allocation in geographically distributed and heterogeneous continuum computing infrastructures, where combinatorial explosion and limited generalization hinder effective deployment. To tackle this, the study introduces, for the first time, the pricing structures commonly found in Software-as-a-Service (SaaS) ecosystems into the resource allocation problem, formulating a unified, price-based representation of the configuration space. The authors propose PRIME, a pricing-aware analysis engine that efficiently searches for cost-optimal deployment configurations satisfying both functional and non-functional constraints. Leveraging synthetic infrastructure topologies and workload generation techniques, the project constructs a comprehensive dataset comprising 9,600 diverse scenarios, demonstrating that the proposed approach achieves both scalability and computational efficiency in complex, heterogeneous environments.

This study investigates the feasibility of achieving fully energy-autonomous task scheduling given energy forecasts, battery capacity constraints, and energy-consuming tasks with time windows. Focusing on task flexibility—defined as the number of time steps within which a task may be scheduled—the work combines computational complexity analysis, fixed-parameter algorithm design, and integer linear programming modeling. It establishes that the problem becomes NP-hard even when flexibility is as low as two, yet remains polynomial-time solvable under certain structural conditions despite high flexibility. Furthermore, the problem is shown to be fixed-parameter tractable with respect to parameters such as the number of tasks. Experimental results demonstrate that moderately increasing task flexibility substantially reduces reliance on external energy sources while maintaining manageable computational overhead.