Capacity-Constrained Influence Maximization: Algorithms and Applications
The paper introduces Capacity‑Constrained Influence Maximization (CIM), a framework that selects up to k neighbors per active user to maximize spread under node capacity limits, proposes MG‑Greedy and RR‑Greedy algorithms with ≥½ approximation, and demonstrates the near‑linear RR‑OPIM+ method’s superior accuracy and speed on large social networks and a Tencent game recommendation system.