Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Local Ratio based Real-time Job Offloading and Resource Allocation in Mobile Edge Computing

Chuanchao Gao, Arvind Easwaran

TL;DR

This work tackles the joint problem of deadline-constrained job offloading and resource allocation in Mobile Edge Computing (MEC), formulating it as the Deadline-constrained Offloading and Resource Allocation Problem ($\mathbf{P}$). It introduces an approximation algorithm, $\mathtt{IDAssign}$, based on a local ratio framework and an instance-dividing technique, and proves a $\frac{1}{6}$-approximation guarantee. The authors model MEC with multiple edge servers, bandwidth and computing resources, and network rings to capture variable channel gains, enabling a uniform ILP formulation that enumerates all feasible assignment instances. Empirical evaluation on real taxi traces and object-detection workloads shows that $\mathtt{IDAssign}$ achieves practical performance close to or exceeding baselines while providing a theoretical bound and scalable runtime. The results highlight the method’s potential for real-time MEC resource management under dual bandwidth and compute constraints, with implications for IoV applications and future online deployments.

Abstract

Mobile Edge Computing (MEC) has emerged as a promising paradigm enabling vehicles to handle computation-intensive and time-sensitive applications for intelligent transportation. Due to the limited resources in MEC, effective resource management is crucial for improving system performance. While existing studies mostly focus on the job offloading problem and assume that job resource demands are fixed and given apriori, the joint consideration of job offloading (selecting the edge server for each job) and resource allocation (determining the bandwidth and computation resources for offloading and processing) remains underexplored. This paper addresses the joint problem for deadline-constrained jobs in MEC with both communication and computation resource constraints, aiming to maximize the total utility gained from jobs. To tackle this problem, we propose an approximation algorithm, $\mathtt{IDAssign}$, with an approximation bound of $\frac{1}{6}$, and experimentally evaluate the performance of $\mathtt{IDAssign}$ by comparing it to state-of-the-art heuristics using a real-world taxi trace and object detection applications.

Local Ratio based Real-time Job Offloading and Resource Allocation in Mobile Edge Computing

TL;DR

This work tackles the joint problem of deadline-constrained job offloading and resource allocation in Mobile Edge Computing (MEC), formulating it as the Deadline-constrained Offloading and Resource Allocation Problem (). It introduces an approximation algorithm, , based on a local ratio framework and an instance-dividing technique, and proves a -approximation guarantee. The authors model MEC with multiple edge servers, bandwidth and computing resources, and network rings to capture variable channel gains, enabling a uniform ILP formulation that enumerates all feasible assignment instances. Empirical evaluation on real taxi traces and object-detection workloads shows that achieves practical performance close to or exceeding baselines while providing a theoretical bound and scalable runtime. The results highlight the method’s potential for real-time MEC resource management under dual bandwidth and compute constraints, with implications for IoV applications and future online deployments.

Abstract

Mobile Edge Computing (MEC) has emerged as a promising paradigm enabling vehicles to handle computation-intensive and time-sensitive applications for intelligent transportation. Due to the limited resources in MEC, effective resource management is crucial for improving system performance. While existing studies mostly focus on the job offloading problem and assume that job resource demands are fixed and given apriori, the joint consideration of job offloading (selecting the edge server for each job) and resource allocation (determining the bandwidth and computation resources for offloading and processing) remains underexplored. This paper addresses the joint problem for deadline-constrained jobs in MEC with both communication and computation resource constraints, aiming to maximize the total utility gained from jobs. To tackle this problem, we propose an approximation algorithm, , with an approximation bound of , and experimentally evaluate the performance of by comparing it to state-of-the-art heuristics using a real-world taxi trace and object detection applications.

Paper Structure

This paper contains 9 sections, 1 theorem, 3 equations, 4 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. System Architecture
  4. Problem Formulation
  5. Instance Dividing Assignment Algorithm
  6. Numerical Evaluation
  7. Simulation Setup
  8. Result Discussion
  9. Conclusion

Key Result

Theorem 1

$\mathtt{IDAssign}$ is a $\frac{1}{6}$-approximation algorithm for $\mathbf{P}$.

Figures (4)

  • Figure 1: (a) An example of mobile edge computing for IoV; (b) Partition the communication range of a wireless network into three network rings;
  • Figure 2: (a) average performance ratios of algorithms under the (low,low) range combination for different jobset sizes; (b) average performance ratios of algorithms under the (high,high) range combination for different jobset sizes; (c) average runtime of algorithms for different jobset sizes (the plots for $\mathtt{Greedy}$ and $\mathtt{Iterative}$ are overlapped);
  • Figure 3: Performance ratios of algorithms under different resource utilization range combinations ($ru_b, ru_c$)
  • Figure 4: average runtime of algorithms for different MEC scales (with jobset size $280$)

Theorems & Definitions (3)

  • Definition 1: Assignment Instance
  • Definition 2: Feasible Assignment Instance
  • Theorem 1