All-pay Auction Based Profit Maximization in End-to-End Computation Offloading System
Hai Xue, Yun Xia, Di Zhang, Honghua Wei, Xiaolong Xu
TL;DR
The paper tackles incentive-compatible pricing for end-to-end MEC offloading by formulating an all-pay auction-based equilibrium pricing scheme. It adopts a SIPV-based three-layer model (EC, aggregation, EU), derives the EU equilibrium bid function $b_{all}(t)$ and the optimal EC reservation value $r^*$, and introduces a minimum valuation and set-based allocation to ensure service access and prevent overload. Key analytical results include the closed-form expressions for $b_{all}(t)$ and the condition $v_0 F(r^*)=\frac{n}{n-\lambda}(n-1)(1-F(r^*)) r^*$, along with a uniform-valuation special case that yields explicit formulas for $b_{all}$ and $r^*$. The proposed scheme outperforms baseline schemes in total profit and achieves lower winner bids, demonstrating improved incentive alignment and efficiency for MEC offloading with guaranteed service access in simulations.
Abstract
Pricing is an important issue in mobile edge computing. How to appropriately determine the bid of end user (EU) is an incentive factor for edge cloud (EC) to offer service. In this letter, we propose an equilibrium pricing scheme based on the all-pay auction model in end-to-end collaboration environment, wherein all EUs can acquire the service at a lower price than the own value of the required resource. In addition, we propose a set allocation algorithm to divide all the bidders into different sets according to the price, and the EUs in each set get the service, which averts the case of getting no service due to the low price. Extensive simulation results demonstrate that the proposed scheme can effectively maximize the total profit of the edge offloading system, and guarantee all EUs can access the service.