Skip to content

Instantly share code, notes, and snippets.

Created October 30, 2018 06:55
What would you like to do?
Code shared from the Rust Playground
use std::sync::{Arc, Mutex, MutexGuard};
use std::thread;
use std::thread::JoinHandle;
type SafeNode = Arc<Mutex<TreeNode>>;
struct TreeNode {
value: u16,
children: Vec<TreeNode>,
fn main() {
let mut root = TreeNode {
value: 1,
children: vec![TreeNode {
value: 2,
children: vec![],
for _i in 1..5 { // for 5 turns, deepen search tree and pick best move
let temp = root;
root = find_best_child(temp);
println!("{:#?}", root);
fn find_best_child(root: TreeNode) -> TreeNode {
let new_children = spawn_mutate_threads(root.children);
.map(|c| {
.expect("unwraping arc")
.expect("unwrapping mutex")
.max_by_key(|c| {
println!("c {:?}", c);
.expect("max by") // now that we've deepend the tree and propogated the node values up, pick the best child node (move) and return it as the new position
fn spawn_mutate_threads(children: Vec<TreeNode>) -> Vec<SafeNode> {
.map(|child| thread(Arc::new(Mutex::new(child)))) // for each possible move, spawn a thread that deepens the search tree
.map(|th| th.join().expect("panicked joining threads"))
fn thread(child: SafeNode) -> JoinHandle<SafeNode> {
thread::spawn(move || {
let child_guard = child.lock().unwrap();
fn mutate(mut child: MutexGuard<TreeNode>) {
for _i in 1..5 {
child.children.push(TreeNode { //simulate deeping search tree by adding new node
value: rand::random(),
children: vec![TreeNode {
value: rand::random(),
children: vec![],
child.value = rand::random(); // value updated to reflect what we learned about the value of this move by evaluating new children
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment