利用改进的并行粒子群算法对变分资料同化的研究  

Research on variational data assimilation based on improved parallel particle swarm optimization

作　　者：董怡琦 周明睿 刘力源 余一冬 陈科 童亚拉[1,2] DONG Yiqi;ZHOU Mingrui;LIU Liyuan;YU Yidong;CHEN Ke;TONG Yala(School of Science, Hubei University of Technology, Wuhan 430068, China;Hubei Engineering Technology Research Center of Energy Photoelectric Device and System, Wuhan 430068, China)

机构地区：[1]湖北工业大学理学院,武汉430068 [2]湖北省能源光电器件与系统工程技术研究中心,武汉430068

出　　处：《华中师范大学学报（自然科学版）》2021年第1期46-51,共6页Journal of Central China Normal University：Natural Sciences

基　　金：湖北省教育厅教学科研项目(2016294,2017320);湖北省教育厅人文社会科学研究项目(17D033);大学生创新创业训练计划项目(国家级)(20191050013).

摘　　要：近年来,为了提高同化精度和减少同化时间,粒子群算法(PSO)被引入到数值天气预报资料同化中来.粒子群算法虽然令同化精度有所提高,但同化时间仍然存在较大缺陷.基于此,首先设计了一种改进的并行粒子群算法(P2PSO),然后应用于含不连续“开关”过程的变分资料同化中,与时变双重压缩因子粒子群算法(PSOTVCF)和动态权重粒子群算法(PSODIWAF)在同化速度、同化精度和收敛性上进行了比较.实验结果表明,设计的并行粒子群算法在不降低同化精度的同时,将同化时间缩短了一半,在收敛速度上明显优于动态权重粒子群算法和时变双重压缩因子粒子群算法.In recent years,in order to improve assimilation precision and reduce assimilation time,Particle Swarm Optimization(PSO)algorithm has been introduced in data assimilation in numerical weather prediction.Although the convergence accuracy has been improved,the assimilation time still has defects.To solve the question,an improved Parallel Particle Swarm Optimization algorithm(P2PSO)was designed firstly,and then it was applied to variational data assimilation with discontinuous“on-off”process.Compared with Particle Swarm Optimization with Dynamic Inertia Weight and Acceleration Factor(PSODIWAF)and Particle Swarm Optimizer with Time Varying Constrict Factor(PSOTVCF)in assimilation speed,assimilation accuracy and convergence,the experimental results show that the designed improved Parallel Particle Swarm Optimization(P2PSO)algorithm reduces the assimilation time by half while maintaining certain advantages in convergence accuracy,and is obviously superior to PSODIWAF and PSOTVCF in the convergence speed.

关 键 词：变分资料同化 数值天气预报 粒子群算法 并行算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]