Distributed/Parallel Genetic Algorithm for Road Traffic Network Division Using Step Parallelization

In this paper, we present the step parallelization of a dividing genetic algorithm (DGA) for the parallel, distributed, and distributed/parallel computing environments. The DGA is employed in a method for road traffic network division, which we developed. The step parallelization, which performs all steps of the genetic algorithm at least partially concurrently, is an alternative to the commonly used island model for the parallelization of genetic algorithms. All three non-sequential executions of the DGA were thoroughly tested and compared to the sequential DGA. The results are also part of the paper.