Evaluating different concurrency configurations for executing multi-agent systems

Reactiveness and performance are important features of Multi-Agent Systems (MAS) and the underlying concurrency model can have a direct impact on them. In multicore programming it is interesting to exploit all the computer cores in order to improve these desirable features. In this paper we perform an experiment to evaluate different concurrency configurations that can be adopted to run an MAS and analyse the effect caused by each configuration on variables like deliberation time and response time. As a result, we identify the advantages and disadvantages for each configuration allowing thus an MAS developer to choose a suitable configuration depending upon the priorities of the application.