Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Joint Edge Server Deployment and Computation Offloading: A Multi-Timescale Stochastic Programming Framework

Huaizhe Liu, Jiaqi Wu, Zhizongkai Wang, Bin Cao, Lin Gao

TL;DR

The paper addresses the challenge of jointly deploying edge servers and offloading computation in MEC when information is realized at different times, introducing a hybrid-information, multi-timescale stochastic programming framework with a strategic layer for ES deployment and a tactical layer for service placement and task offloading. It develops a Lyapunov-based online algorithm (SPCO) for the tactical layer and a Markov-approximation-based algorithm (MAIED) for the strategic layer, providing theoretical bounds on optimality gaps and complexity. Simulation results show SP-JESO outperforming baselines with up to 56% total cost reduction, validating the benefits of exploiting partial information and timescale separation. The work offers practical insight for deploying MEC in B5G/6G scenarios with DIAs, where upfront infrastructure decisions must align with dynamic runtime optimization.

Abstract

Mobile Edge Computing (MEC) is a promising approach for enhancing the quality-of-service (QoS) of AI-enabled applications in the B5G/6G era, by bringing computation capability closer to end-users at the network edge. In this work, we investigate the joint optimization of edge server (ES) deployment, service placement, and computation task offloading under the stochastic information scenario. Traditional approaches often treat these decisions as equal, disregarding the differences in information realization. However, in practice, the ES deployment decision must be made in advance and remain unchanged, prior to the complete realization of information, whereas the decisions regarding service placement and computation task offloading can be made and adjusted in real-time after information is fully realized. To address such temporal coupling between decisions and information realization, we introduce the stochastic programming (SP) framework, which involves a strategic-layer for deciding ES deployment based on (incomplete) stochastic information and a tactical-layer for deciding service placement and task offloading based on complete information realization. The problem is challenging due to the different timescales of two layers' decisions. To overcome this challenge, we propose a multi-timescale SP framework, which includes a large timescale (called period) for strategic-layer decision-making and a small timescale (called slot) for tactical-layer decision making. Moreover, we design a Lyapunov-based algorithm to solve the tactical-layer problem at each time slot, and a Markov approximation algorithm to solve the strategic-layer problem in every time period.

Joint Edge Server Deployment and Computation Offloading: A Multi-Timescale Stochastic Programming Framework

TL;DR

The paper addresses the challenge of jointly deploying edge servers and offloading computation in MEC when information is realized at different times, introducing a hybrid-information, multi-timescale stochastic programming framework with a strategic layer for ES deployment and a tactical layer for service placement and task offloading. It develops a Lyapunov-based online algorithm (SPCO) for the tactical layer and a Markov-approximation-based algorithm (MAIED) for the strategic layer, providing theoretical bounds on optimality gaps and complexity. Simulation results show SP-JESO outperforming baselines with up to 56% total cost reduction, validating the benefits of exploiting partial information and timescale separation. The work offers practical insight for deploying MEC in B5G/6G scenarios with DIAs, where upfront infrastructure decisions must align with dynamic runtime optimization.

Abstract

Mobile Edge Computing (MEC) is a promising approach for enhancing the quality-of-service (QoS) of AI-enabled applications in the B5G/6G era, by bringing computation capability closer to end-users at the network edge. In this work, we investigate the joint optimization of edge server (ES) deployment, service placement, and computation task offloading under the stochastic information scenario. Traditional approaches often treat these decisions as equal, disregarding the differences in information realization. However, in practice, the ES deployment decision must be made in advance and remain unchanged, prior to the complete realization of information, whereas the decisions regarding service placement and computation task offloading can be made and adjusted in real-time after information is fully realized. To address such temporal coupling between decisions and information realization, we introduce the stochastic programming (SP) framework, which involves a strategic-layer for deciding ES deployment based on (incomplete) stochastic information and a tactical-layer for deciding service placement and task offloading based on complete information realization. The problem is challenging due to the different timescales of two layers' decisions. To overcome this challenge, we propose a multi-timescale SP framework, which includes a large timescale (called period) for strategic-layer decision-making and a small timescale (called slot) for tactical-layer decision making. Moreover, we design a Lyapunov-based algorithm to solve the tactical-layer problem at each time slot, and a Markov approximation algorithm to solve the strategic-layer problem in every time period.

Paper Structure

This paper contains 49 sections, 3 theorems, 43 equations, 13 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Background and Motivations
  3. Solution Approach and Contributions
  4. Related Works
  5. Joint Optimization under Complete Information
  6. Joint Optimization under Stochastic Information
  7. Novelty of Our Work
  8. System Model
  9. Network Model
  10. Application and Service Model
  11. Computation Model
  12. Communication Model
  13. Communication links between UEs and ES
  14. Communiation links between ES
  15. Stochastic Information Model
  16. ...and 34 more sections

Key Result

Theorem 1

The time-average tactical-layer cost achieved by the SPCO algorithm satisfy the condition: where $\gamma^\circ$ denotes the offline optimal tactical-layer cost under complete information, and $B$ is the constant in que-upper-bound.

Figures (13)

  • Figure 1: Decision-making based on (a) complete information, (b) incomplete information, and (c) hybrid information.
  • Figure 2: An example of MEC network with DIA applications.
  • Figure 3: Decision-Making based on hybrid information in multi-period and multi-slot scenario.
  • Figure 4: Time-average Performance of (a) Tactical-layer cost (b) Service operation cost (c) UE delay cost.
  • Figure 5: Convergence properties of (a) The MAIED algorithm (b) The total system cost dynamic with $\beta-V$
  • ...and 8 more figures

Theorems & Definitions (3)

  • Theorem 1: Optimality
  • Theorem 2: Queue Stability
  • Theorem 3: Optimality Gap of MAIED Algorithm