






摘" 要: 為了讓用戶更快地獲取內容,提出一種基于競價模式和智能迭代模式的綜合式邊緣協作緩存算法。在競價算法部分,采用熵權法計算中間值、訪問熱度、緩存空間和內容熱度的權重,并綜合計算節點上內容的緩存分數,根據緩存分數進行緩存替換。在迭代算法部分,將網絡空間劃分為多個子域,每個子域通過混合遺傳算法進行周期性計算,并得出子域內內容緩存方案。實驗結果表明,所提算法在降低云服務器負載、減少請求平均跳數方面具有明顯的優勢。
關鍵詞: 邊緣計算; 協作緩存; 競價模式; 智能迭代模式; 熵權法; 網絡空間; 混合遺傳算法
中圖分類號: TN919?34; TP311" " " " " " " " " "文獻標識碼: A" " " " " " " " " " " 文章編號: 1004?373X(2024)12?0165?05
Research on edge collaborative caching algorithm based on bidding mode
and intelligent iteration mode
LI Changminchen1, LU Feipeng2, WANG Yanling2
(1. Hangzhou Dianzi University, Hangzhou 310018, China; 2. Hangzhou Newgrand Technology Co., Ltd., Hangzhou 310000, China)
Abstract: In order to enable users to obtain content faster, a comprehensive edge collaborative caching algorithm based on bidding mode and intelligent iteration mode is proposed. In the bidding algorithm section, the entropy weight method is used to calculate the weights of intermediary value, access heat, cache space, and content heat. The cache score of the content on the node is calculated comprehensively, and cache replacement is performed based on the cache score. In the iterative algorithm section, the network space is divided into multiple subdomains, each of which is periodically computed by means of the hybrid genetic algorithm, and a content caching scheme within the subdomains is obtained. The experimental results show that the proposed algorithm has significant advantages in reducing cloud server load and average hop count of requests.
Keywords: edge computing; collaborative caching; bidding mode; intelligent iteration mode; entropy weight method; cyberspace; hybrid genetic algorithm
0" 引" 言
隨著各行業數字化改造的持續推進,越來越多的終端設備接入網絡,導致網絡中的數據流量呈現井噴式增長[1]。然而,龐大的數據流量極易引發網絡擁塞甚至堵塞,從而增加用戶接入時延,并嚴重降低用戶服務體驗。因此,在智能交通、智能醫療和智能工業等物聯網環境下對設備終端的數據緩存研究尤為關鍵[2]。傳統云計算架構一般遠離大部分用戶地理位置,因此會影響服務質量和用戶體驗[3]。針對上述問題,邊緣計算應運而生。邊緣計算緩存是指將數據、應用或計算結果存儲在邊緣計算節點上,以提高數據訪問速度并降低網絡延時和帶寬消耗[4]。通過將數據緩存到離用戶更近的地方,從而減少數據在網絡上傳輸的距離,并提升數據訪問速度和響應時間?!?br>