Analysis of All to All Broadcast on multi mesh of trees using genetic algorithm

Various communication algorithms have been proposed on a variety of network architectures. However, an efficient algorithm on an efficient architecture is always the focus point. We have proposed all to all broadcast (AAB) algorithm on multi mesh trees using parallel random access machine (PRAM), which combines the topological properties of multi-mesh and mesh of trees architecture. The algorithm is elaborated by taking eight processors in each mesh. Further, the proposed algorithm has been scaled by the genetic algorithm (GA) approach (for correctness) that passes the algorithm through functions such as fitness, crossover, mutation and evaluation. The fitness function deals with conflicts that occur during communication. Crossover and mutation look for new solutions by computing probability of crossover (p_cc) and probability of mutation (p_mm) that are compared by an evaluation function.