A Heterogeneous Multiprocessor Independent Task Scheduling Algorithm Based on Improved PSO

The independent task scheduling problem of heterogeneous multiprocessors belongs to the NP-hard problem. The emergence of evolutionary algorithms provides a new idea for solving this problem. Particle swarm optimization (PSO) is a kind of intelligent evolutionary algorithm and it could be used to solve scheduling problem. We firstly discretized the representation of particle swarm optimization algorithm and made it suitable for the scheduling problem of heterogeneous multiprocessors. Then, the PSO algorithm was introduced into heterogeneous multiprocessors independent task scheduling problem by modeling method. In order to overcome particle swarm optimization algorithm's problem that is easy to fall into local optimum and premature convergence. We proposed a heterogeneous multiprocessor independent task scheduling algorithm based on improved PSO by improving the update operation of particle swarm optimization algorithm and transformed it into crossover and mutation operation of genetic algorithm. The experimental results show that the improved PSO scheduling algorithm can overcome the premature defects of PSO algorithm and the makespan of proposed IPSO is smaller than PSO.