Extremal coalitions for influence games through swarm intelligence-based methods
| Autor: | Maria Serna, Fabi醤 Riquelme, Rodrigo Olivares, Xavier Molinero, Francisco Mu駉z |
|---|---|
| Přispěvatelé: | Universitat Politècnica de Catalunya. Departament de Matemàtiques, Universitat Politècnica de Catalunya. Departament de Ciències de la Computació, Universitat Politècnica de Catalunya. GRTJ - Grup de Recerca en Teoria de Jocs, Universitat Politècnica de Catalunya. ALBCOM - Algorismia, Bioinformàtica, Complexitat i Mètodes Formals |
| Jazyk: | angličtina |
| Rok vydání: | 2022 |
| Předmět: |
Influence spread
Computer science business.industry Collective behavior Swarm intelligence Influence game Computer Science Applications Biomaterials Mechanics of Materials Modeling and Simulation Bio-inspired computing Artificial intelligence Electrical and Electronic Engineering Jocs Teoria de business Matemàtiques i estadística::Investigació operativa::Teoria de jocs [Àrees temàtiques de la UPC] Game theory |
| Zdroj: | UPCommons. Portal del coneixement obert de la UPC Universitat Politècnica de Catalunya (UPC) |
| Popis: | An influence game is a simple game represented over an influence graph (i.e., a labeled, weighted graph) on which the influence spread phenomenon is exerted. Influence games allow applying different properties and parameters coming from cooperative game theory to the contexts of social network analysis, decision-systems, voting systems, and collective behavior. The exact calculation of several of these properties and parameters is computationally hard, even for a small number of players. Two examples of these parameters are the length and the width of a game. The length of a game is the size of its smaller winning coalition, while the width of a game is the size of its larger losing coalition. Both parameters are relevant to know the levels of difficulty in reaching agreements in collective decision-making systems. Despite the above, new bio-inspired metaheuristic algorithms have recently been developed to solve the NP-hard influence maximization problem in an efficient and approximate way, being able to find small winning coalitions that maximize the influence spread within an influence graph. In this article, we apply some variations of this solution to find extreme winning and losing coalitions, and thus efficient approximate solutions for the length and the width of influence games. As a case study, we consider two real social networks, one formed by the 58 members of the European Union Council under nice voting rules, and the other formed by the 705 members of the European Parliament, connected by political affinity. Results are promising and show that it is feasible to generate approximate solutions for the length and width parameters of influence games, in reduced solving time "Funding Statement: F. Riquelme has been partially supported by Fondecyt de Iniciación 11200113, Chile, and by the SEGIB scholarship of Fundación Carolina, Spain; X. Molinero under grants PID2019-104987GB-I00 (JUVOCO); M. Serna under grants PID2020-112581GB-C21 (MOTION) and 2017-SGR-786 (ALBCOM)" |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|