Multi-Task Combinatorial Bandits for Budget Allocation

TL;DR

This paper introduces a multi-task combinatorial bandit framework for optimizing budget allocation across multiple advertising campaigns, leveraging Bayesian models and Thompson sampling to enhance decision-making under uncertainty.

Contribution

It presents a novel online system that combines hierarchical Bayesian modeling with flexible techniques and Thompson sampling for efficient budget allocation in advertising.

Findings

System effectively maximizes cumulative returns in experiments.

Demonstrates robustness and adaptability across different environments.

Integrates diverse modeling techniques for complex scenarios.

Abstract

Today's top advertisers typically manage hundreds of campaigns simultaneously and consistently launch new ones throughout the year. A crucial challenge for marketing managers is determining the optimal allocation of limited budgets across various ad lines in each campaign to maximize cumulative returns, especially given the huge uncertainty in return outcomes. In this paper, we propose to formulate budget allocation as a multi-task combinatorial bandit problem and introduce a novel online budget allocation system. The proposed system: i) integrates a Bayesian hierarchical model to intelligently utilize the metadata of campaigns and ad lines and budget size, ensuring efficient information sharing; ii) provides the flexibility to incorporate diverse modeling techniques such as Linear Regression, Gaussian Processes, and Neural Networks, catering to diverse environmental complexities; and iii) employs the Thompson sampling (TS) technique to strike a balance between exploration and exploitation. Through offline evaluation and online experiments, our system demonstrates robustness and adaptability, effectively maximizing the overall cumulative returns. A Python implementation of the proposed procedure is available at https://anonymous.4open.science/r/MCMAB.