最不经常使用LFU(least frequently used)算法java实现

最不经常使用LFU(least frequently used)算法java实现

LFU(least frequently used (LFU) page-replacement algorithm)。即最不经常使用页置换算法,要求在页置换时置换引用计数最小的页,因为经常使用的页应该有一个较大的引用次数。但是有些页在开始时使用次数很多,但以后就不再使用,这类页将会长时间留在内存中,因此可以将引用计数寄存器定时右移一位,形成指数衰减的平均使用次数。

以上是来自百度百科的介绍,通俗来说,LFU 就是一个淘汰算法,例如内存相较于硬盘可以快速的读写,但是内存的空间是有限的,为了有效的利用内存空间,就需要将一些非热点数据给淘汰掉,这里就用到 LFU 算法,它会将内存中最不经常使用的数据给淘汰掉。

LFU 算是 LRU(Least Recently Used) 的一种变种算法,LRU 算法在找工作面试中,经常会被问到,需要能够在短时间内写出来,大家平时还是要勤加练习。

LRU 算法的 java 算法实现请参见我另一篇博文:最近最少使用LRU(Least Recently Used)算法java实现

动动发财小手,关注 + 点赞 + 收藏不迷路。

一.手撸 LFU 算法实现(Hash表 + 双向链表)

import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.atomic.AtomicInteger;

/**
 * LFU (Least frequently used)最不经常使用
 *
 * @Date: 2022/01/07 10:10
 */
public class LfuCache {

    private static ConcurrentHashMap<String, ListNode> cacheMap;

    private static AtomicInteger count;

    private static int capacity;

    private final ListNode head = new ListNode();

    private final ListNode tail = new ListNode();

    LfuCache(int capacity) {
        count = new AtomicInteger(0);
        LfuCache.capacity = capacity;
        cacheMap = new ConcurrentHashMap<>();
        this.tail.pre = head;
        this.head.next = tail;
    }

    private static class ListNode {

        ListNode() {
        }

        ListNode(String key) {
            this.key = key;
            this.count = 1;
        }

        private String key;

        private int count;

        private ListNode pre;

        private ListNode next;

        public String getKey() {
            return this.key;
        }

    }

    public void put(ListNode node) {
        ListNode existNode = get(node);
        if (existNode != null) {
            // cacheMap中node存在,先remove节点,再把节点移到链表头部
            existNode.count++;
            ListNode replaceNode = existNode.pre;
            while (replaceNode != head && replaceNode.count < existNode.count) {
                replaceNode = replaceNode.pre;
            }
            if (replaceNode == head) {
                removeNode(existNode);
                headAdd(existNode);
            }
            // 即:replaceNode.count >= existNode.count,在链表中移除existNode,并移动到replaceNode的后面
            replaceNode = replaceNode.next;
            if (replaceNode != tail && replaceNode != existNode) {
                removeNode(existNode);
                existNode.next = replaceNode;
                existNode.pre = replaceNode.pre;
                replaceNode.pre.next = existNode;
                replaceNode.pre = existNode;
            }
        } else {
            // cacheMap中node不存在,插入
            if (count.get() == capacity) {
                tailRemove();
                // count--
                count.decrementAndGet();
            }
            tailAdd(node);
            // count++
            count.incrementAndGet();
            cacheMap.put(node.getKey(), node);
        }
    }

    public void headAdd(ListNode node) {
        node.next = head.next;
        head.next.pre = node;
        node.pre = head;
        head.next = node;
    }

    public void tailAdd(ListNode node) {
        tail.pre.next = node;
        node.pre = tail.pre;
        node.next = tail;
        tail.pre = node;
    }

    public void removeNode(ListNode node) {
        node.pre.next = node.next;
        node.next.pre = node.pre;
    }

    public void tailRemove() {
        tail.pre.pre.next = tail;
        tail.pre = tail.pre.pre;
    }

    public ListNode get(ListNode node) {
        return cacheMap.get(node.getKey());
    }

    public void print() {
        ListNode node = head;
        System.out.println("============================================");
        while (node.next != null && node.next != tail) {
            System.out.println(node.next.key + ", key count = " + node.next.count + ", LruCache count = " + count);
            node = node.next;
        }
    }

    public void print(String key) {
        ListNode node = head;
        System.out.println("================  put " + key + "  ================");
        while (node.next != null && node.next != tail) {
            System.out.println(node.next.key + ", key count = " + node.next.count + ", LruCache count = " + count);
            node = node.next;
        }
    }

    public static void main(String[] args) {
        ListNode node1 = new ListNode("1");
        ListNode node2 = new ListNode("2");
        ListNode node3 = new ListNode("3");
        ListNode node4 = new ListNode("4");
        LfuCache lruCache1 = new LfuCache(3);
        lruCache1.put(node1);
        lruCache1.print();
        lruCache1.put(node2);
        lruCache1.print();
        lruCache1.put(node3);
        lruCache1.print();
        lruCache1.put(node4);
        lruCache1.print();

        System.out.println("");
        System.out.println("###################################################");
        System.out.println("");

        ListNode node9 = new ListNode("9");
        LfuCache lruCache2 = new LfuCache(3);
        lruCache2.put(node9);
        lruCache2.print();
        lruCache2.put(node9);
        lruCache2.print();

        System.out.println("");
        System.out.println("###################################################");
        System.out.println("");

        ListNode node5 = new ListNode("5");
        ListNode node6 = new ListNode("6");
        ListNode node7 = new ListNode("7");
        ListNode node8 = new ListNode("8");
        LfuCache lruCache3 = new LfuCache(4);
        lruCache3.put(node5);
        lruCache3.print("5");
        lruCache3.put(node6);
        lruCache3.print("6");
        lruCache3.put(node7);
        lruCache3.print("7");
        lruCache3.put(node5);
        lruCache3.print("5");
        lruCache3.put(node5);
        lruCache3.print("5");
        lruCache3.put(node8);
        lruCache3.print("8");
        lruCache3.put(node8);
        lruCache3.print("8");

        System.out.println("");
        System.out.println("###################################################");
        System.out.println("");

        ListNode node10 = new ListNode("10");
        ListNode node11 = new ListNode("11");
        ListNode node12 = new ListNode("12");
        ListNode node13 = new ListNode("13");
        LfuCache lruCache4 = new LfuCache(3);
        lruCache4.put(node10);
        lruCache4.put(node10);
        lruCache4.put(node10);
        lruCache4.print();
        lruCache4.put(node11);
        lruCache4.print();
        lruCache4.put(node12);
        lruCache4.put(node12);
        lruCache4.print();
        lruCache4.put(node13);
        lruCache4.print();
    }

}

输出如下:

============================================
1, key count = 1, LruCache count = 1
============================================
1, key count = 1, LruCache count = 2
2, key count = 1, LruCache count = 2
============================================
1, key count = 1, LruCache count = 3
2, key count = 1, LruCache count = 3
3, key count = 1, LruCache count = 3
============================================
1, key count = 1, LruCache count = 3
2, key count = 1, LruCache count = 3
4, key count = 1, LruCache count = 3

###################################################

============================================
9, key count = 1, LruCache count = 1
============================================
9, key count = 2, LruCache count = 1

###################################################

================  put 5  ================
5, key count = 1, LruCache count = 1
================  put 6  ================
5, key count = 1, LruCache count = 2
6, key count = 1, LruCache count = 2
================  put 7  ================
5, key count = 1, LruCache count = 3
6, key count = 1, LruCache count = 3
7, key count = 1, LruCache count = 3
================  put 5  ================
5, key count = 2, LruCache count = 3
6, key count = 1, LruCache count = 3
7, key count = 1, LruCache count = 3
================  put 5  ================
5, key count = 3, LruCache count = 3
6, key count = 1, LruCache count = 3
7, key count = 1, LruCache count = 3
================  put 8  ================
5, key count = 3, LruCache count = 4
6, key count = 1, LruCache count = 4
7, key count = 1, LruCache count = 4
8, key count = 1, LruCache count = 4
================  put 8  ================
5, key count = 3, LruCache count = 4
8, key count = 2, LruCache count = 4
6, key count = 1, LruCache count = 4
7, key count = 1, LruCache count = 4

###################################################

============================================
10, key count = 3, LruCache count = 1
============================================
10, key count = 3, LruCache count = 2
11, key count = 1, LruCache count = 2
============================================
10, key count = 3, LruCache count = 3
12, key count = 2, LruCache count = 3
11, key count = 1, LruCache count = 3
============================================
10, key count = 3, LruCache count = 3
12, key count = 2, LruCache count = 3
13, key count = 1, LruCache count = 3

Process finished with exit code 0

二.总结

本文是在 最近最少使用LRU(Least Recently Used)算法java实现 基础上的变种算法,有兴趣的可以参考一下。

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值