Best Poster Award at ACM MobiHoc 2025 for SeLR: Sparsity-enhanced Lagrangian Relaxation for Edge Computation Offloading
Published:
Sparsity-enhanced Lagrangian Relaxation (SeLR) for Computation Offloading at the Edge
(Negar Erfaniantaghvayi, Zhongyuan Zhao, Kevin Chan, Ananthram Swami, Santiago Segarra)
We are delighted to share that our poster “Sparsity-enhanced Lagrangian Relaxation (SeLR) for Computation Offloading at the Edge” received the Best Poster Award at ACM MobiHoc 2025, held in Houston, TX.
Congratulations to Negar Erfaniantaghvayi, who led this work and is about to defend her PhD — an outstanding achievement recognized by the MobiHoc community! 🎉
This work introduces SeLR, a convex relaxation framework that combines primal–dual optimization with reweighted ℓ₁-regularization to tackle the NP-hard problem of joint task offloading and routing in edge networks.
Compared to traditional greedy heuristics, SeLR significantly reduces the optimality gap while maintaining comparable runtime, offering a practical balance between scalability and near-optimality for real-time decision-making at the edge.
The poster version is available on ACM DL, and the extended paper can be found on arXiv.
Figure 1. Best Poster Award at ACM MobiHoc 2025 for SeLR.