Abstract: To solve the problem of particle swarm optimization algorithm, which is easy to fall into local optimum in network on chip mapping, a hybrid optimization mapping Algorithm based on particle swarm optimization and genetic algorithm was proposed. It implemented GA and PSO separately with two groups, by the superior individuals from GA algorithm instead of the initial random particles from PSO algorithm, which not only maintained the diversity of the group but also improved search efficiency. Simulation results based on NS-2 show that the Network-on-Chip from the automatic generation tools have a good performance in network latency, throughput, and link bandwidth optimization comparing the random mapping under the same amount of computation scale.

Key words: 2D Mesh, hybrid algorithm, NoC mapping, NS-2