Title: MB-CIM: a multi-round budgeted competitive influence maximisation

Authors: Nadia Niknami; Jie Wu

Addresses: Department of Computer and Information Sciences, Temple University, Philadelphia, PA, USA ' Department of Computer and Information Sciences, Temple University, Philadelphia, PA, USA

Abstract: Influence maximisation is trying to select a small set of seed users in the social network to maximise the spread of influence. An individual's decision to adopt a product or innovation will be highly dependent on the choices made by the individual's neighbours in the social network. In CIM, competitors need to decide which nodes would be an influential one and how many resources should be allocated to this member. Identifying the best algorithm for the influence maximisation under budget constraint has become a demanding task. Predicting, and controlling social influence is an exciting topic in social network analysis. We are interested in multi-round CIM where each competitor needs to decide the location and the amount of budget to invest in the most influential members simultaneously and repeatedly under a given total budget. This paper proposes a tree-approximate game-theoretical framework and introduces the new measurement as a dynamic node weight.

Keywords: budget allocation; game theory; reinforcement learning; social networks; multi-round influence maximisation.

DOI: 10.1504/IJSN.2023.129896

International Journal of Security and Networks, 2023 Vol.18 No.1, pp.30 - 41

Received: 20 Jan 2022
Accepted: 25 Jan 2022
Published online: 03 Apr 2023 *