张晓丽.基于学习速率与更新向量的混合云数据冗余值迭代算法[J].河南理工大学学报（自然科学版）,2020,39(5):114-119.

ZHANG X L.Iterative algorithm for redundancy value of hybrid cloud data based onlearning rate and update vector[J].Journal of Henan Polytechnic University(Natural Science) ,2020,39(5):114-119.

基于学习速率与更新向量的混合云数据冗余值迭代算法

张晓丽1,2

1.东北大学机械工程与自动化学院，辽宁 沈阳 110819;2.长春师范大学数学学院，吉林 长春 130032

摘要:针对传统混合云数据冗余值迭代算法的平均回报值较低、收敛稳定性较差、收敛动作规模较小等问题，提出一种基于学习速率与更新向量的混合云数据冗余值迭代算法。首先，构建混合云数据冗余值值函数，在该函数中引入一个新的参数更新权重向量，基于深度学习中学习速率要求，获取值函数的稳定值；其次，依据获取的稳定值计算值函数稳定值向量，利用新权值处理稳定值向量，获取值函数更新向量；最后，对权值增量进行计算，结合哈希表完成混合云数据冗余值的迭代研究。实验结果表明，该算法的平均回报值最高，且收敛速度最快。

关键词:学习速率;更新向量;深度学习;混合云数据;冗余值迭代算法

doi:10.16186/j.cnki.1673-9787.2020.5.17

基金项目:国家自然科学基金资助项目（41071262，61300230）

收稿日期:2020/01/09

修回日期:2020/03/10

出版日期:2020/09/15

Iterative algorithm for redundancy value of hybrid cloud data based onlearning rate and update vector

ZHANG Xiaoli1,2

1.School of Mechanical Engineering and Automation， Northeastern University， Shenyang  110819 ， Liaoning， China;2.School of Mathematics， Changchun Normal University， Changchun  130032 ， Jilin， China

Abstract:Aiming at the problems of low average return value， poor convergence stability and small scale of convergence action for the traditional hybrid cloud data redundancy value iterative algorithm， a hybrid cloud data redundancy value iterative algorithm based on learning rate and update vector was proposed. Firstly， a hybrid cloud data redundancy value function was constructed， a new parameter was introduced into the function to update the weight vector， and a stable value of the value function was acquired based on learning rate requirements of deep learning ； Secondly， the value function was calculated according to the obtained stable value， the stable value vector was used to process the stable value vector with new weights， and the value function update vector was obtained； Finally， the weight increment was calculated， the iterative study of the redundant value of hybrid cloud data was completed by combining the Hash table with the calculation. The experimental results showed that the average return value of the algorithm was the highest and the convergence speed was the fastest.

Key words:learning rate;update vector;deep learning;hybrid cloud data;redundancy value iteration algorithm

基于学习速率与更新向量的混合云数据冗余值迭代算法_张晓丽.pdf