OPTIMIZING THE DR-SUBMODULAR FUNCTION ON THE INTEGER LATTICE TO MAXIMIZE THE INFLUENCE OF VIRAL MARKETING ON COMMUNITIES

Abstract

As society continues to develop, individuals encounter increasingly complex optimization problems with multiple objectives to achieve. One such problem is optimizing a DR-submodular function, characterized by diminishing returns. In this article, we focus on maximizing the influence of marketing spread on social network communities, using a novel technique known as streaming data browsing. Our proposed DR-SubOptStream algorithm yields positive results, achieving an acceptable approximation of the objective function value and better complexity than existing algorithms. To conduct experiments, we transform social network data from a connected graph form to bipartite data form and then run the algorithm on preprocessed datasets. Overall, our findings demonstrate the effectiveness of our approach in solving this type of problem.