The Development of the agent-based demography and migration model of Eurasia and its supercomputer implementation

Abstract

In this work we describe the development of a scalable agent-based modelling framework for simulation of Eurasia population described in terms of demography, migration and transport flows. The simulated system will consist of agents representing individuals and sets of links to other agents, which represent the social interactions of individual. The individual agents in the model will participate in several independent processes, for which different sets of social links is important such as family and neighbors. As a base for our simulation system we have used a combination of a base native layer implemented using C++ language which uses MPI library, and Microsoft .NET platform as an environment for model code written in high-level C# programming language. To perform a load balancing of agents between processes the METIS/ParMETIS algorithms were used. These algorithms allow to split the graph of agents and links into parts of similar size with the least possible number of links between them. A number of numerical experiments were carried out for test model to estimate the influence of the parameters of the model on its performance and parallel scalability. For each combination of parameters a number of simulations were performed to average the results.