Simulation on task scheduling for multiprocessors based on improved neural network

This paper mainly discusses task scheduling for multiprocessors. Application requires higher performance of the multiprocessors task scheduling systems. The traditional algorithms majorly consider the accuracy and neglect the real-time performance. In order to improve the real-time performance while maintaining the accuracy, the paper proposes a task scheduling algorithm (GA-ACO) for multiprocessors based on improved neural network. It first builds mathematical models for task scheduling of multiprocessor systems, and then introduces genetic algorithms to quickly find feasible solutions. The simulation results show that the improved neural network algorithm not only has the global optimization ability of genetic algorithm, but also has both local search and the positive feedback capabilities of neural networks; compared with single optimization algorithm, it can quickly find the task scheduling solutions to meet real-time requirements, accelerate the speed of execution of the task, furthermore achieve reasonable, effective task allocation and scheduling for multi-processor.