A Time Performance Evaluation of the Soma Asynchronous Parallel Distribution in Java and C#

This paper compares two different implementations of the Self-Organizing Migrating Algorithm (SOMA), which is a highly effective tool of an evolutionary optimization that is aimed at the same set of problems as Genetic Algorithms. One implementation of algorithm was created in the C# framework and the second implementation in Java framework. Both implementations are the asynchronous parallel ‘All-to-One’ strategy of SOMA, which is used to equally distribute computation loads between several available processors/cores. The aim of our effort is to statistically evaluate the computation time efficiency of these two concurrent frameworks including dependence on the number of threads. The obtained results are discussed in the conclusion.