Abstract
A parallel bacterial foraging algorithm was developed for the multiple sequence alignment problem. Four sets of homologous genetic and protein sequences related to Alzheimer's disease among various species were collected from the NCBI database for convergence analysis and performance comparison. The main question was the following: is the bacterial foraging algorithm suitable for the multiple sequence alignment problem? Three versions of the algorithm were contrasted by performing a t-test and Mann-Whitney test based on the results of a 30-run scheme, focusing on fitness, execution time, and the number of function evaluations as performance metrics. Additionally, we conducted a performance comparison of the developed algorithm with the well-known Genetic Algorithm. The results demonstrated the consistent efficiency of the bacterial foraging algorithm, while the version of the algorithm based on gap deletion presented an increased number of function evaluations and excessive execution time. Overall, the first version of the developed algorithm was found to outperform the second version, based on its efficiency. Finally, we found that the third bacterial foraging algorithm version outperformed the Genetic Algorithm in the third phase of the experiment. The sequence sets, the algorithm's Python 3.12 code and pseudocode, the data collected from the executions, and a GIF animation of the convergence on various different sets are available for download.