EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

A New Framework for Guaranteed Click Ad Allocation in Mobile Location-Based Advertising Networks

Ngoc-Bach Pham, Tien-Dung Nguyen, The-Phong Le, Anh Son Ta

IEEE Access · 2026 · 14 · 61923-61936

DOI: 10.1109/access.2026.3684915 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Ngoc-Bach Pham und 3 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: AWING Technology & Media JSC.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Guaranteed Click (GC) advertising allows advertisers to secure a predetermined number of clicks for targeted audiences while providing publishers with predictable revenue. Delivering these campaigns optimally is challenging due to budget, pacing, targeting, and inventory constraints. The difficulty is amplified in mobile location-based settings, such as Wi-Fi marketing networks, where inventory supply varies across hotspots and user experience constraints limit the frequency of ad displays per location. We propose a two-stage optimization framework for GC ad allocation in location-based advertising environments. The first stage computes a fair baseline allocation that respects campaign demand and inventory limits while minimizing weighted deviations from campaign-specific target shares. The second stage then maximizes a revenue-based objective subject to the same demand and inventory constraints, plus an equity floor that guarantees each campaign at least its baseline exposure. To solve the second-stage problem at scale, we introduce a Pure Lagrangian Relaxation with Projection (PLRP) method that enables closed-form primal updates, and uses projected dual ascent with a hybrid stepsize policy: a constant stepsize for the demand multipliers ( $\alpha $ ) enabled by L-smoothness of the partial dual, and a conservative diminishing stepsize for the supply multipliers ( $\beta $ ) to ensure convergence when smoothness with respect to $\beta $ is not guaranteed under the projection-based primal oracle. We show that, under standard convexity assumptions, PLRP converges to a Karush-Kuhn-Tucker (KKT) point of the second-stage problem. Through simulations and offline log-replay experiments on anonymized production logs from a commercial Wi-Fi hotspot advertising network, we demonstrate that our framework achieves near-optimal fulfillment of guaranteed contracts while maintaining high click-through rates (CTR) and respecting location-level delivery caps. Compared with state-of-the-art baselines, PLRP consistently improves both delivery and user engagement across a range of demand and supply scenarios, and yields more equitable exposure for budget-constrained advertisers such as small and medium-sized businesses.

Förderung

  • AWING Technology & Media JSC

Zitieren

APA

Pham, N., Nguyen, T., Le, T., & Ta, A. S. (2026). A New Framework for Guaranteed Click Ad Allocation in Mobile Location-Based Advertising Networks. IEEE Access, 14, 61923-61936. https://doi.org/10.1109/access.2026.3684915

Vancouver

Pham N, Nguyen T, Le T, Ta AS. A New Framework for Guaranteed Click Ad Allocation in Mobile Location-Based Advertising Networks. IEEE Access. 2026;14:61923-61936. doi:10.1109/access.2026.3684915

BibTeX
@article{Pham2026,
  author = {Pham, Ngoc-Bach and Nguyen, Tien-Dung and Le, The-Phong and Ta, Anh Son},
  title = {A New Framework for Guaranteed Click Ad Allocation in Mobile Location-Based Advertising Networks},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {61923--61936},
  doi = {10.1109/access.2026.3684915}
}
Kostenlos starten →