Improved Particle Optimization Algorithm Solving Hadoop Task Scheduling Problem

Cloud computing to provide service for the user group is huge, so the number of cloud computer's tasks is enormous, the system handle large tasks all the time so that task scheduling is the key and difficult points in the cloud. This article make research on how to make full use of cloud resources for task efficiently scheduling. This paper proposes an Improved Particle Swarm-Estimation of Distribution optimization Algorithm (IPS-EDA) based on task allocation strategy. The task scheduling strategy is optimization strategy based on improved particle swarm algorithm, which introduce estimation of distribution algorithm (EDA) based probabilistic model and random sampling theory, the proposed algorithm does not fall into local optimum. The simulation results show that the performance of IPS-EDA has been greatly improved provides better load balancing and resource utilization.