Simulating information diffusion in a multidimensional social network using the DEVS formalism (WIP) Youssef Bouanan 1 , Mathilde Forestier 1 , Judicael Ribault 1 , Gregory Zacharewicz 1 , Bruno Vallespir 1 , Nejib Moalla 2 1 Univ. Bordeaux, IMS, UMR 5218, F-33400 Talence, France. 2 Univ. Lyon, LIESP, EA 4125, F-69500, Bron, France. 1 {firstname.lastname}@ims-bordeaux.fr, 2 nejib.moalla@univ-lyon2.fr ABSTRACT The impact of information on individuals within a social network is, mostly, statically modeled and the dynamic is not frequently tackled. In addition, the work of modeling and simulation of the population’s reactions to the information do not use explicit specification languages to describe their models. These models are specified in the shape of graph or math formulas and then directly implemented and coded using classical programming languages. We propose to model the actions of influence in a multidimensional social network (MSN). Each graph layer corresponds to a predetermined social network based on one relationship. In this work, the use of the DEVS formalism has permitted to explicit M&S of human behavior and the interaction between individuals as a network. In more detail, we define a set of models of individuals characterized by a set of state variables (e.g., using Maslow’s theory [15] to construct the behavior of an individual) and the mesh between the individuals within a social network. Then, we introduce the platform architecture, sharing resources, specifically designed to simulate MSN. In the end, a scenario is used to validate our models using the platform based on DEVS Specification. Author Keywords Multidimensional social network; DEVS Formalism; Information Propagation; Human behavior; Modeling and simulation. ACM Classification Keywords I.6 SIMULATION AND MODELING (e.g. Applications). : Miscellaneous INTRODUCTION AND MOTIVATION Human behavior can be difficult to understand and predict, thus it can be qualified as a complex system. DEVS is a well-defined formalism, which has numerous advantages over other formalisms in the modeling of complex dynamic systems. The purpose of this work is to provide a simple but efficient and accurate framework to model the behavior of an individual, but also to simulate the propagation of information among a group of individuals and its influence on their behavior. General definitions define human behavior as the collection of behaviors demonstrated by humans [14]. Behaviors are influenced by numerous aspects (e.g., culture, attitudes, emotions, values, ethics, authority, rapport, persuasion, coercion, etc.). Also, humans have many ways to communicate (e.g. word of mouth, phone, SMS, emails, and the mass diffusion as radio or TV). The communication takes place in the social networks where the individuals are involved. We propose a greatly simplified model of human behavior and the message dissemination in the social networks using a defined media. Then these models are validated by simulation. The communication will be established thanks to the individual connection with other individuals within the different social networks. The results will consist of measuring the diffusion of the information and the ability to reach the targeted people. In literature most multidimensional social networks (MSN) are flattened at the implementation of the solution. Today some approaches formalize MSN but they are little used in computer practice. No complete implementation is done that integrates the social networks and the dynamic message propagation. Most current MSN-based simulations flatten the different networks into one, which serves to manage all the network- specific rules into one place. This approach makes it hard to develop, validate and, ultimately, reuse the model. A shared component as proposed in [5], offers a good opportunity to have one human behavior model shared in several networks. DEVS is a timed, highly modular, hierarchical formalism for the description of reactive systems. It can be appropriated to implement networks, propagation and human behavior. A few related works have provided DEVS models of human behavior that we will use with slight modifications; Seck et al. present a DEVS based framework for the modelling and simulation of human behavior with the influence of stress and fatigue [20]. Faucher et al. proposed a first approach using G-DEVS formalism for Civil-Military Cooperation actions (CIMIC) and Psychological actions (PSYOPS), which are actions of influence that take precedence over combat [7]. In more detail, this paper will participate in the definition of a set of models that addresses the entities and the structure of a population working in firms. It will begin by TMS/DEVS 2015, April 12 - 15, 2015, Alexandria, VA, USA © 2015 Society for Modeling & Simulation International (SCS)