基于竞争遗传算法的海洋云平台资源调度模型研究

高质量的海洋自然资源管理离不开数据和信息的支撑。鉴于海洋数据的特殊性，海洋数据处理常涉及长时间序列或大空间范围的处理工作，对于此类密集型计算为主的数据处理，通用型云平台存在效率不高的突出问题。文章在全面分析Hadoop平台原生资源调度算法的基础上，结合海洋数据处理密集型计算的特点，创新性地提出了基于竞争模型的遗传算法任务调度策略(CGA),有效地解决了遗传算法求解速度受初始化种群与种群进化测量影响较大的问题。此外，为加快收敛速度，引入竞争机制，构建基于种群竞争的自适应进化模型。通过实际验证和比对，证明改进后的算法在收敛速度及收敛结果的稳定性上都优于传统算法，有效地改进了海洋云平台资源调度的能力，提升了海洋数据的处理效率。

Research on Resource Scheduling Model of Ocean Cloud Platform Based on Competitive Genetic Algorithm

High quality marine natural resource management cannot be achieved without the support of data and information. Given the unique nature of marine data, the processing of marine environmental data often involves long time series or large-scale processing work. For data processing work mainly focused on intensive computing, universal cloud platforms have prominent issues of low efficiency. Based on a comprehensive analysis of the native resource scheduling algorithms on the Hadoop platform and the characteristics of intensive computing in ocean data processing, this paper innovatively proposes a genetic algorithm task scheduling strategy based on a competitive model. The use of chaotic algorithm mechanism as the basis for genetic algorithm population initialization ensures that the solution space obtained during each initialization can be uniformly distributed, effectively solving the problem of genetic algorithm solving speed being greatly affected by the initialization population and population evolution measurement. In addition, in order to speed up rate of convergence, competition mechanism is introduced and an adaptive evolution model based on population competition is proposed. Through the actual verification and comparison of the built models, it is proved that the improved algorithm is superior to the traditional genetic algorithm in rate of convergence and stability of the convergence results, and has greatly improved the ability and efficiency of improving the resource scheduling of the ocean cloud platform.