Stochastic dynamic programming heuristics for influence maximization-revenue optimization

摘要

The well-known influence maximization (IM) problem has been actively studied by researchers over the past decade with emphasis on marketing and social networks. Existing research has determined solutions to the IM problem by obtaining the influence spread and utilizing the property of submodularity. This paper is based on a novel approach to the IM problem geared toward optimizing clicks, and consequently revenue, within an online social network (OSN). Our approach differs from existing approaches by adopting a novel, decision-making perspective using stochastic dynamic programming (SDP). Thus, we define a new problem, influence maximization-revenue optimization and propose SDP as a solution method. The SDP method has lucrative gains for an advertiser in terms of optimizing clicks and generating revenue, but one drawback to the method is its associated "curse of dimensionality" particularly for large OSNs. Thus, we introduce the L-degree heuristic, adaptive hill-climbing heuristic and the multistage particle swarm optimization heuristic as methods which are orders of magnitude faster than the SDP method while achieving near-optimal results. We compare these heuristics on both synthetic and real-world networks.