Adaptive Budget Optimization for Multichannel Advertising Using Combinatorial Bandits

2025-02-05

AI Summary

This paper addresses the dynamic budget allocation problem for multi-channel digital advertising under non-stationary environments. We propose an adaptive composite bandit algorithm that jointly incorporates saturation response modeling and change-point detection to enable targeted exploration and rapid adaptation to environmental shifts. We develop the first high-fidelity, open-source simulation environment supporting long-horizon, multi-channel advertising campaigns. Additionally, we introduce a domain-knowledge-driven target-region filtering strategy to enhance decision-making efficiency. Theoretically, we derive a regret upper bound for the proposed method. Empirical evaluation across multiple real-world advertising campaigns demonstrates an average 12.7% improvement in cumulative reward and a 34.5% reduction in cumulative regret compared to state-of-the-art baselines, confirming its superior performance and robustness.

Technology Category

• Machine Learning: Online Learning & Bandits
• Search and Optimization: Sampling/Simulation-based Search
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Effective budget allocation is crucial for optimizing the performance of digital advertising campaigns. However, the development of practical budget allocation algorithms remain limited, primarily due to the lack of public datasets and comprehensive simulation environments capable of verifying the intricacies of real-world advertising. While multi-armed bandit (MAB) algorithms have been extensively studied, their efficacy diminishes in non-stationary environments where quick adaptation to changing market dynamics is essential. In this paper, we advance the field of budget allocation in digital advertising by introducing three key contributions. First, we develop a simulation environment designed to mimic multichannel advertising campaigns over extended time horizons, incorporating logged real-world data. Second, we propose an enhanced combinatorial bandit budget allocation strategy that leverages a saturating mean function and a targeted exploration mechanism with change-point detection. This approach dynamically adapts to changing market conditions, improving allocation efficiency by filtering target regions based on domain knowledge. Finally, we present both theoretical analysis and empirical results, demonstrating that our method consistently outperforms baseline strategies, achieving higher rewards and lower regret across multiple real-world campaigns.

Problem

• Optimize budget allocation in digital advertising
• Adapt to non-stationary market dynamics
• Develop practical simulation and allocation algorithms

Innovation

• Simulates multichannel advertising campaigns
• Uses combinatorial bandit with saturating mean
• Incorporates change-point detection for adaptation