16.25. LRU cache

16. Moderate problems/16.25. lru_cache.md

@@ -0,0 +1,58 @@
+# Least Recently Used cache
+
+> Design and build a LRU cache, which evicts the least recently used item. The cache should map from keys to values (allowing to insert and retrieve a value associated with a particular key), and be initialized with a max size. When it is full, it should evict the LRU item.
+
+## First idea
+
+Create linked list with max size, add in forward order. When a value is to be retrieved, visit the node and put it in the back. Keep a counter for number of items added, when it surpasses `max_len`, drop the head.
+
+```java
+class Node {
+    Node next = null;
+    int key;
+    int value;
+    int max_size;
+
+    public Node(int k, int v, int max_size){
+        this.key = k;
+        this.value = v;
+        this.max_size = max_size;
+    }
+
+    void appendToTail(int k, int v){
+        Node end = new Node(k, v);
+        Node head_tmp = this;
+        Node n = this;
+
+        int i = 0;
+        while (n.next != null){
+            // if max value is achieved, delete head node (the one accessed the last)
+            if(i >= head_tmp.max_size){
+                head_tmp = head_tmp.next;
+                i--;
+            }
+            n = n.next;
+            i++;
+        }
+        n.next = end;
+    }
+
+    int retrieve_value(int k){
+        while (n.next != null){
+            if(n.key == k){
+                return n.value;
+            }
+            n = n.next;
+        }
+        return -1
+    }
+}
+```
+
+## Hints
+
+Both a hash table and a doubly linked list would be useful, combine the two.
+
+## Solution
+
+A singly linked list is fast for eviction but slow for retrieval of keys. Make a doubly linked list, where it's easy to remove the middle element. The hash table maps to each linked list node rather than the value.