Heterogeneous platform task scheduling method and system based on Q learning

Abstract

The invention belongs to the technical field of heterogeneous multi-processor computing, and particularly relates to a heterogeneous platform task scheduling method and system based on Q learning, which take all tasks as a Q learning state space, take a processor set as an action space, and wait for an allocated task as a current state. The method comprises the following steps: obtaining a task initial mapping scheme according to execution time required for mapping a task to an action space in Q learning; creating a genetic algorithm model, carrying out fitness evaluation on the task initial mapping scheme, setting individuals copied to the next generation of population in the genetic algorithm model according to fitness, carrying out crossover variation on reserved individuals, and determining new population optimization efficiency and a minimum threshold value; obtaining an approximately optimal solution mapped from a task to a processor in the model according to the genetic algorithm model; converting the model approximate optimal solution into ant colony information initial information distribution, and iteratively searching and outputting an optimal path through an ant colonyalgorithm according to the information distribution to obtain a task scheduling optimal scheme, so as to better improve the performance of the heterogeneous platform.

Description

technical field [0001] The invention belongs to the technical field of heterogeneous multiprocessor computing, and in particular relates to a Q-learning-based heterogeneous platform task scheduling method and system. Background technique [0002] With the continuous improvement of high-performance computing requirements for various signal processing tasks and the rapid development of hardware accelerators, general-purpose processors have been unable to meet the needs of strong real-time and large-scale computing, and heterogeneous computing systems are increasingly used to solve Complex task handling problems. The heterogeneous architecture includes a series of processors with very different structures such as CPU, GPU, FPGA, and DSP. The processors are connected through special networks or interfaces to meet the hardware performance requirements of different types of computing tasks. Improve resource utilization and computing efficiency. To meet the demands of increasingl...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problems of low flexibility, slow convergence speed, and poor predictability in existing scheduling algorithms, the present invention provides a heterogeneous platform task scheduling method and system based on Q-learning, which can adjust the network search direction in time while taking into account Local and global search to obtain better results, maximize the efficiency of each processor on the heterogeneous platform, facilitate parallel processing of tasks, and improve the performance of the heterogeneous platform

Images