Research On Task Scheduling Strategy Based On The Improved Fireworks Algorithm

In recent years,computer architecture has continued to develop,the variety and number of multi-core processors have gradually increased,and the structure of various types of processors has become increasingly complex.How to effectively exploit the parallel performance of multi-core processors has been a hot research topic in the computer field.The task scheduling strategy of multi-core processors directly affects the parallel performance of multi-core systems and the efficiency of processing application.The optimal scheduling sequence of poor task scheduling algorithm is not high precision,the task scheduling time is long,and the utilization of processor core is low.To improve the parallel capability of multi-core processors and improve the operating efficiency of multi-core systems,it is necessary to design an efficient task scheduling strategy.The paper has analyzed the research status of multi-core processors task scheduling problem,and deeply studied the architecture and technical characteristics of multi-core processors.At the same time,the working principle and performance characteristics of fireworks algorithm has been analyzed.Combined with the characteristics of multi-core processors task scheduling problem,a task scheduling algorithm based on improved fireworks algorithm-IGFWA has been proposed.The IGFWA algorithm map the continuous solution space of the fireworks algorithm to the discrete solution space of the multi-core processors task scheduling problem,so that the fireworks algorithm can solve the task scheduling problem;the exclusion operator has been added in the explosion process of the fireworks,and the Euclidean distance threshold has been set.When the distance between the sparks and the fireworks was too close,the effectiveness of the sparks was ensured by the repulsion operation to avoid invalid sparks due to the excessively small explosion,which wastes computational resources.At the same time,due to the poor effect of the Gaussian variation process,the population cannot be effectively increased.Diversity,and easy to cause the convergence of the algorithm to slow down,so the gene position mutation strategy has been used instead of Gaussian mutation to ensure the mutation process was effective.In addition,the selection strategy based on fitness value has been used instead of the Euclidean distance-based selection strategy to improve the algorithm.The operating efficiency has been obtained in a short time to obtain an optimal task scheduling sequence.To verify that IGFWA can effectively solve the multi-core processor task scheduling problem and test the performance of IGFWA algorithm,the IGFWA algorithm has been compared with the basic fireworks algorithm and genetic algorithm in Matlab platform.According to the experimental results,the IGFWA algorithm can converge faster than the others.The solution has higher precision and can obtain a better task scheduling sequence in a shorter time.