An Adaptive Genetic Algorithm for Mesh-Based NoC Application Mapping

Abstract

Application mapping is one of the key problems of Network-on-Chip (NoC) design. It maps the cores of application to the processing elements of the NoC topology. This paper presents a novel approach for NoC application mapping, which uses adaptive genetic algorithm (AGA) in the mapping. The proposed approach adaptively varies the probabilities of crossover and mutation operators in genetic algorithm, aiming to reduce the overall communication cost of NoC. Experimental results show that the proposed approach decreases the communication cost by 3% to 7% on average, compared to the existing approach using Standard Genetic Algorithm (SGA).