Shapley Value Framework for Fairness in Budgeted Combinatorial Bandits

A new research paper proposes a meritocratic fairness framework for budgeted combinatorial multi-armed bandits with full-bandit feedback (BCMAB-FBF). The authors extend the Shapley value from cooperative game theory to a K-Shapley value, which measures marginal contributions of agents restricted to sets of size at most K. They prove that K-Shapley value uniquely satisfies symmetry, linearity, null player, and efficiency properties. The paper introduces K-SVFair-FBF, a fairness-aware bandit algorithm that adaptively estimates K-Shapley values under unknown valuation functions, addressing the challenge of full-bandit feedback where individual arm contributions are not directly observed. This work advances fairness in sequential decision-making with budget constraints.