Parallel-EvoCluster: An Open-Source Parallel Nature-Inspired Optimization Clustering

Bio-inspired algorithms are a branch of Artificial Intelligence in which the behavior of natural systems is emulated to design non-deterministic heuristic methods for search, clustering, optimization, and simulation. Clustering algorithms search for patterns in the data and identify groups into clusters having similar features. The objective of this research is to improve the execution times of ten bio-inspired serial algorithms to solve clustering problems. To achieve our objective, we use parallelization techniques to reduce execution time by utilizing all processing resources. We proposed Parallel-EvoCluster to parallelize ten algorithms through two parallelization approaches, such as shared-memory and distributed-memory architecture using Open Multi-Processing and Message Passing Interface APIs respectively. Both approaches were performed using 24 CPU cores. The results show a significant improvement in runtimes of the parallel algorithms, parallelization with MPI on average is 10.14x times faster and the average parallelization with OMP is 9.05x times faster than the serial versions. The code is available at: https://github.com/win7/Parallel_EvoCluster .