[Paper Review] Game-based Pricing and Task Offloading in Mobile Edge Computing Enabled Edge-Cloud Systems

Yi Su, Wenhao Fan+2 more|arXiv (Cornell University)|Jan 14, 2021

IoT and Edge/Fog Computing25 references4 citations

TL;DR

This paper proposes a distributed game-theoretic mechanism for task offloading and pricing in mobile edge computing (MEC)-enabled edge-cloud systems, using a multi-leader multi-follower Stackelberg game to model interactions between offloading service providers (OSPs) and IoT mobile devices (MDs). It introduces two algorithms—Iterative Proximal Offloading Algorithm (IPOA) for MDs and Iterative Stackelberg Pricing Algorithm (ISPA) for OSPs—demonstrating that IPOA reduces MD disutility significantly and ISPA boosts OSP revenue with a price of anarchy below 150%.

ABSTRACT

As a momentous enabling of the Internet of things (IoT), mobile edge computing (MEC) provides IoT mobile devices (MD) with powerful external computing and storage resources. However, a mechanism addressing distributed task offloading and price competition for the open exchange marketplace has not been established properly, which has become a huge obstacle to MEC's application in the IoT market. In this paper, we formulate a distributed mechanism to analyze the interaction between OSPs and IoT MDs in the MEC enabled edge-cloud system by appling multi-leader multi-follower two-tier Stackelberg game theory. We first prove the existence of the Stackelberg equilibrium, and then we propose two distributed algorithms, namely iterative proximal offloading algorithm (IPOA) and iterative Stackelberg game pricing algorithm (ISPA). The IPOA solves the follower non-cooperative game among IoT MDs and ISPA uses backward induction to deal with the price competition among OSPs. Experimental results show that IPOA can markedly reduce the disutility of IoT MDs compared with other traditional task offloading schemes and the price of anarchy is always less than 150\%. Besides, results also demonstrate that ISPA is reliable in boosting the revenue of OSPs.

Motivation & Objective

• Address the lack of a distributed, competitive mechanism for task offloading and pricing in MEC-enabled IoT systems without a central authority.
• Model the interaction between multiple selfish OSPs (leaders) and IoT MDs (followers) in a decentralized marketplace.
• Formulate a disutility function for MDs that jointly captures delay, energy consumption, and payment costs to reflect quality of experience (QoE).
• Design a distributed algorithm (IPOA) for MDs to find Nash equilibrium offloading strategies under fixed OSP prices.
• Develop a dynamic pricing algorithm (ISPA) using backward induction to enable OSPs to iteratively adjust prices based on non-cooperative competition.

Proposed method

• Formulate a multi-leader multi-follower two-tier Stackelberg game to model OSP-MD interactions in a competitive edge-cloud environment.
• Define a disutility function for MDs that integrates queuing delay, transmission energy, and monetary cost to quantify QoE.
• Propose the Iterative Proximal Offloading Algorithm (IPOA) to solve the follower-level non-cooperative game among MDs using proximal point methods.
• Introduce the Iterative Stackelberg Pricing Algorithm (ISPA) to solve the leader-level non-cooperative game among OSPs via backward induction.
• Use approximation of utility derivatives in ISPA to enable dynamic, privacy-preserving price updates without full knowledge of MD utility functions.
• Apply backward induction to compute equilibrium prices iteratively, ensuring convergence to a Stackelberg equilibrium.

Experimental results

Research questions

• RQ1 Does a Stackelberg equilibrium exist in a multi-leader multi-follower game model of task offloading and pricing in MEC-enabled edge-cloud systems?
• RQ2 Can a distributed algorithm (IPOA) effectively compute Nash equilibrium offloading strategies for IoT MDs under fixed OSP prices?
• RQ3 How can OSPs dynamically adjust their prices in a non-cooperative, privacy-preserving manner to maximize their utility?
• RQ4 To what extent does the proposed mechanism improve system efficiency and fairness compared to traditional offloading schemes?
• RQ5 What is the performance of the mechanism in terms of disutility reduction for MDs and revenue enhancement for OSPs?

Key findings

• The proposed mechanism achieves a price of anarchy below 150%, indicating that the system performance loss due to selfish behavior remains bounded.
• IPOA significantly reduces the disutility of IoT MDs compared to traditional offloading schemes, demonstrating improved QoE.
• ISPA successfully increases the revenue of OSPs, with a 7.5% improvement in average utility over blind pricing after 50 iterations.
• OSP prices increase rapidly with iteration count and are higher for edge computing providers than cloud providers due to their performance advantage.
• The utility of all OSPs increases with each iteration of ISPA, and edge computing OSPs consistently achieve higher utility than cloud computing OSPs.
• The average disutility of MDs increases with higher transmission power (εᵢᵗˣ), but IPOA maintains low disutility across varying task arrival rates and power levels.