基于GA-DNN模型的RRAM数据自动化分析系统  

Automation Analysis System for RRAM Data Based on GA-DNN Model

作　　者：王晓荃 左石凯 陈铖颖 WANG Xiaoquan;ZUO Shikai;CHEN Chengying(School of Optoelectronic&Communication Engineering,Xiamen University of Technology,Xiamen Fujian 361024,China)

机构地区：[1]厦门理工学院光电与通信工程学院,福建厦门361024

出　　处：《电子器件》2024年第5期1201-1209,共9页Chinese Journal of Electron Devices

基　　金：福建省自然科学基金面上项目(2020J01295);厦门市青年创新基金项目(3502Z20206074);厦门市重大科技项目(3502Z20221022)。

摘　　要：为了整合RRAM器件的测试数据,提升RRAM器件性能探知效率,构建了一套适用于RRAM测试数据分析和可视化的Web系统。结合Set和Reset操作的测试数据,提出一种GA-DNN模型算法,用于预测RRAM器件操作电压,从而为相同测试条件的实验提供可信的预测范围。相比于BPNN模型和CNN模型,在测试集相对误差小于15%的评估条件下,GA-DNN模型使得Set电压的预测精度分别提升了11.38%和2.33%;同时,Reset电压的预测精度分别提升了9.25%和2.54%。最终,系统采用B/S架构、Vue.js前端框架和Django后端框架建立了前后端分离的RRAM器件测试数据分析系统。该系统可进行RRAM器件的测试数据统计、可视化分析以及测试电压预测。In order to integrate test data and improve the efficiency of performance detection, a Web system for resistive random access memory(RRAM)test data analysis and visualization is built. Combining with the test data of Set and Reset operations, a genetic algorithm deep neural networks(GA-DNN) model algorithm is proposed, which is used to predict the operating voltage and provide a credible range of the same test conditions. Compared with the BPNN and CNN model, the GA-DNN model improves the prediction accuracy of the Set voltage by 11.38% and 2.33% respectively when the relative error of the test set is less than 15%,and the accuracy is improved by 9.25% and 2.54% respectively for Reset voltage. Finally, the system adopts B/S architecture, Vue.js front-end and Django back-end framework to establish a test data analysis system with separate front and back ends. It can execute test data statistics, visual analysis and test voltages prediction.

关 键 词：阻变存储器 GA-DNN模型 自动化分析系统 数据分析及可视化 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]