Business Dynamics
Network Analysis: Methodological Foundations (Lecture Notes in Computer Science)
Network Analysis: Methodological Foundations (Lecture Notes in Computer Science)
| Hi-index | 0.00 |
The Paper describes a proposed approach to modelling a wide range of phenomena under a working common name of "spreading of epidemics". Examples of actual processes falling into that kind include spreading of infectious disease within a population (hence the name), spreading of information or rumours within a society, popularisation of a new product (technology, standard), travel of internet worms in computer networks etc. There are several existing models dealing with such processes and a number of simulation tools for conducting numerical experiments, among them CARE (Creative Application to Remedy Epidemics, with which both authors were involved). Although an effort was undertook to implement in CARE most of relevant advances in such modelling, one can point out several deficiencies of that system, such as: assumed network structure of connections among individuals, assumed and forced homogeneity of agents, only one kind of disease spreading at a time, focus on diseases only (without regard to a number of similar phenomena). The proposed approach aims at overcoming these imperfections. Here, we share our concept of constructing an appropriate modelling framework by means of a multi-agent simulation. We use a popular software tool for building and executing multi-agent simulations, called RePast, which we briefly describe. The agents' strategies are initially modelled as FSM and then relatively easily transferred to the format acceptable by RePast. We present a few preliminary ideas for simulation scenarios and agents' strategies. The results obtained from simulation experiments would include both: time-series (and derived statistics) about the spread of the analysed factor within population (numbers of healthy, infected, etc.) but also the effective structure of network of contacts that emerges from the agents' activities. As a side-effect, in a way, the model could be used for verification of results obtained from CARE.