Leveraging Reinforcement Learning with Greedy Path Mining for Target Group Inuence Maximization in Social Networks

Binesh, Neda; Haddadi, Mahdieh

doi:10.22075/ijnaa.2026.37384.5440

Leveraging Reinforcement Learning with Greedy Path Mining for Target Group Inuence Maximization in Social Networks

Articles in Press

Document Type : Research Paper

Authors

Neda Binesh

Mahdieh Haddadi

Faculty of Mathematics, Statistics and Computer Science Semnan University, Semnan, Iran

10.22075/ijnaa.2026.37384.5440

Abstract

Social networks are characterized by the formation of groups, where decisions are often influenced by majority consensus.
This paper deals with the issue of Target Group Influence Maximization (TGIM). The objective of TGIM is to select a set of k influential nodes to maximize the activation of target group members through the dissemination of specific content or information.
We propose a novel algorithm, Target Group Influence Maximization using Reinforcement Learning
with Greedy Path mining (TGRLGP), to address the TGIM problem. TGRLGP leverages reinforcement learning techniques to identify optimal paths from target nodes to potential influencers. Experimental results demonstrate the superior performance of the proposed TGRLGP algorithm over existing methods.
Extensive experiments on four benchmark networks show that TGRLGP consistently achieves the highest influence spread among the compared approaches, improving influence spread by up to 4.76% and by 2.12% on average over the strongest baseline.

Keywords

Q-learning

Information diffusion

Independent cascade

Seed selection

Monte Carlo rollout

Heuristic search

Subjects