基于K-means的动态聚类垃圾回收算法研究

打开文本图片集

中图分类号：TP333；TP301.6 文献标识码：A 文章编号：2096-4706（2025）07-0093-05

Abstract： Inconsumer electronics products based onNAND flash memory，garbage colection significantlyafects device performance，andcold-hot data separation is the keytooptimizing the garbage colection algorithm.This paper proposes an inovative method that uses the K-means clustering algorithm to automaticall determinethedata heat intervals and achieve eficientcold-hotdataseparation.Additionally，anadaptiveifluence factoradjustmentstrategyisdesignedtodyamically balancetheinfuenceofhistoricaldataheatandrecentdataheatOnthisbasis，thegarbagecolectionstrategyisoptimized.The simulationresults showthatcompared withtheexistingalgorithms，this methodhasasignifcant efectonimproving thereadwriteperformanceand achieving wear leveling，which is helpful in prolonging the service lifeoftheequipment and improving the overall performance.

Keywords： flash memory; garbage collection; cold-hot separation; K-means; clustering

0 引言

随着移动设备、嵌入式系统和数据中心的普及，NANDFlash闪存在存储领域的重要性日益凸显。（剩余7766字）