Research On EBBO Task Scheduling Algorithm Based On Chip Multi-processors

With the golden age of modern computer architecture,multi-core architecture has become the choice of more and more processors.However,the current task scheduling algorithms of multi-core processor systems have problems such as slow convergence and difficulty in jumping out of local optimums.Therefore,it is of great application value and practical significance to propose a more stable and efficient task scheduling algorithm that makes full use of the powerful computing power of chip multi-processors and improves the parallelism of applications.The thesis focuses on the shortcomings of biogeographic algorithms and proposes algorithm improvement strategies.The concept of habitat similarity is introduced to characterize the differences between individuals.When initializing the ecosystem,only individuals with a sufficiently small similarity are allowed to join the population,expanding the initial solution set in the solution space.The migration pressure coefficient is used to improve the algorithm migration model.This coefficient is proportional to the number of algorithm iterations.In the later period of the algorithm,the increase of the pressure coefficient can reduce the loss of excellent individual characteristics and reduce the probability of random habitat migration.When the algorithm first converged,the firework algorithm explosion operator was introduced.The three habitats with the best fitness value,the worst,and the center were used as fireworks to explode,perturb it,search the neighborhood solution space,and jump out of the local optimal solution.At the same time,an improved biogeographic algorithm is used to solve the static task scheduling problem of heterogeneous chip multi-processors.The continuous solution space of the biogeography algorithm is transformed into the discrete solution space of task scheduling.A system model is established to represent the calculation rate and Communication rate.A task model is established to represent the amount of computation and communication of the task nodes.At the same time,design a reasonable encoding and decoding method for the habitat,use the reciprocal of the task scheduling length as the fitness index of the algorithm,apply the improved algorithm to the multi-core processor scheduling and update the algorithm initialization method,migrate the mutation model and the explosion operator to form the final Task scheduling strategy EBBO.In order to verify the feasibility and efficiency of the EBBO multi-core processor task scheduling algorithm,a simulation experiment was designed on the Matlab platform.The EBBO task scheduling algorithm was compared with the Ant Colony Optimization,Particle Swarm Optimization and their improved algorithms.Test the algorithm's optimization ability,stability,and convergence speed.The experimental results show that the algorithm is feasible on the static task scheduling problem of heterogeneous chip multi-processors.Compared with the control algorithm,it can obtain a shorter task scheduling sequence in a shorter time and has better stability.