Task matching and scheduling based on co-evolutionary model

Task matching and scheduling plays an important role in parallel and distributed systems. The efficacy and efficiency of conventional single population-based evolutionary algorithms (CEAs) decrease with the number of independent tasks. By analyzing the mechanism that makes CEA in-scalable, this paper proposes a task matching and scheduling algorithm based on the computational model of cooperative co-evolution, which is inspired by the co-evolutionary phenomena of natural species. Then, we discuss some problems related to the proposed algorithm such as construction of initial population, genetic operators including improved crossover and migration as a kind of mutation, cooperative interactions among species and fitness computation of individual. The algorithm was analyzed mathematically, which shows that the exponential increase index of the co-evolution based scheduling algorithm is higher than that of CEA. Simulation results verify the theoretical result