Delay-Optimal Computation Task Scheduling for Mobile-Edge Computing Systems (1604.07525v3)

Abstract: Mobile-edge computing (MEC) emerges as a promising paradigm to improve the quality of computation experience for mobile devices. Nevertheless, the design of computation task scheduling policies for MEC systems inevitably encounters a challenging two-timescale stochastic optimization problem. Specifically, in the larger timescale, whether to execute a task locally at the mobile device or to offload a task to the MEC server for cloud computing should be decided, while in the smaller timescale, the transmission policy for the task input data should adapt to the channel side information. In this paper, we adopt a Markov decision process approach to handle this problem, where the computation tasks are scheduled based on the queueing state of the task buffer, the execution state of the local processing unit, as well as the state of the transmission unit. By analyzing the average delay of each task and the average power consumption at the mobile device, we formulate a power-constrained delay minimization problem, and propose an efficient one-dimensional search algorithm to find the optimal task scheduling policy. Simulation results are provided to demonstrate the capability of the proposed optimal stochastic task scheduling policy in achieving a shorter average execution delay compared to the baseline policies.

• The paper introduces a two-timescale stochastic optimization framework based on a Markov decision process for delay-optimal task scheduling under power constraints.
• It employs a one-dimensional search algorithm to determine optimal offloading decisions, significantly reducing average execution delay compared to conventional methods.
• The study highlights practical benefits for MEC systems by enhancing user experience in delay-sensitive applications such as online gaming and video conferencing.

Delay-Optimal Computation Task Scheduling for Mobile-Edge Computing Systems

The paper "Delay-Optimal Computation Task Scheduling for Mobile-Edge Computing Systems" addresses a critical aspect of mobile-edge computing (MEC): the effective scheduling of computation tasks to optimize delay under power constraints. The authors propose a policy leveraging a two-timescale stochastic optimization framework enabled by a Markov decision process (MDP), focusing on minimizing the average execution delay while considering energy consumption constraints.

Problem Formulation and Methodology

The paper formulates the task scheduling problem in MEC systems as a two-part challenge across different timescales. On a larger timescale, the decision to process tasks locally on mobile devices or offload them to MEC servers is addressed, while on a smaller timescale, channel conditions influence the transmission policy for task input data.

To tackle this, the authors adopt an MDP-based scheduling strategy founded on the current state of the task queue, the execution unit, and the transmission unit. They derive a power-constrained delay minimization problem and propose a one-dimensional search algorithm to ascertain the optimal task scheduling policy.

Numerical Findings

Simulation results, comparing the proposed method against several baseline approaches, underscore the efficacy of the scheduling policy in significantly reducing average execution delay. For example, the methodology successfully manages to capitalize on parallel task processing capabilities inherent in MEC systems, thereby optimizing both local device and MEC server resources.

Additionally, the paper's simulations illustrate the adaptive behavior of the proposed policy under varying task arrival rates and channel conditions, revealing a stable improvement in computation experience compared to alternative policies like local execution or cloud-only approaches.

Implications and Future Directions

The research demonstrates practical implications for advancing MEC systems to better support computation-intensive and delay-sensitive applications. By efficiently managing task scheduling with a robust stochastic approach, there is potential to remarkably enhance the quality of experience (QoE) for end-users, supporting applications such as online gaming and video conferencing.

Theoretically, this work contributes to the body of knowledge on stochastic optimization in edge computing systems. Its application of MDPs in deriving optimal task offloading decisions could inform future research in adapting similar frameworks for more complex MEC ecosystems.

Speculation on Future Developments

Future advancements could explore extending the model to more distributed MEC systems involving multiple mobile devices and servers, potentially encompassing heterogeneous network conditions. Furthermore, integrating adaptive learning mechanisms to refine MDP parameters based on real-time network feedback could yield even more efficient scheduling decisions.

In conclusion, this paper offers a significant step toward optimizing task scheduling in MEC systems, paving the way for further exploration and practical implementation in diverse computing environments.