Empirical analysis of the parallelization of the Monte Carlo simulation and its usage in the field of Economy

Abstract

Monte carlo simulation is one of the techniques used to generate pseudorandomnumbers which can comprise a certain set of random data. Thesedata are used in different fields of study, especially in those which requirelarge amounts of data to predict a given situation, for example the market ineconomy, risk management, particle movement in physics,telecommunication, applied statistics and many others.Since this generation of data requires a lot of time if executed on a singlemachine, the best approach is to use parallel systems. These systemsunfortunately are not used in the albanian reality even though they bring a lotof advantages not only for the academic world, but even to importantcompanies or institutions [2].In our paper we will analyze some of the most important methods applied inthe field of economy. The monte carlo algorithm, as one of the mostimportant and wide used algorithms in economic simulations, will be seen inthe parallelization point of view. We will consider empirical characteristicssuch as speed-up and especially the complexity of this parallelizedalgorithm. By doing so we will demonstrate some of the advantages ofparallel systems such as time and money savings.