Improvement of the Ensemble Empirical Mode Decomposition with Parallel Computing

In this paper we propose a novel approach to increase a processing speed of the Ensemble Empirical Mode Decomposition algorithm and its Complementary version with usage of parallel computations. It is shown that this computational scheme is effective and leads to a significant increase in processing speed up to approximately 4.1 times with 8 parallel branches in comparison with non-parallel algorithm, reducing the needed time for decomposition to the 24% of initial calculations. Further increase in the computational performance of parallel EEMD heavily relies on reducing the parallel overhead costs. Numerical estimations and tests are presented in order to verify and to prove the final results and conclusions, as well as to confirm the theoretical assumptions on the computational complexity of parallel EEMD as well.