Pass explore / exploit tuning param as another scoring config

1 files changed, 37 insertions, 11 deletions

diff --git a/src/strategy/minimax.rs b/src/strategy/minimax.rs
index 55abcec..2c52127 100644
--- a/src/strategy/minimax.rs
+++ b/src/strategy/minimax.rs
@@ -21,6 +21,7 @@ pub struct ScoreConfig {
     victory_weight: f32,
     snowball_weight: f32,
     bomb_weight: f32,
+    explore_exploit_weight: f32,
 }
 
 impl Default for ScoreConfig {
@@ -32,6 +33,7 @@ impl Default for ScoreConfig {
             victory_weight: 3000.,
             snowball_weight: 100.,
             bomb_weight: 100.,
+            explore_exploit_weight: 10.,
         }
     }
 }
@@ -198,7 +200,7 @@ fn expand_tree(node: &mut Node, mut state: GameBoard, config: &ScoreConfig) -> S
 
         score
     } else {
-        let commands = choose_existing(node);
+        let commands = choose_existing(node, config);
         state.simulate(commands);
         let score = expand_tree(
             node.children
@@ -265,21 +267,45 @@ fn score(state: &GameBoard, config: &ScoreConfig) -> Score {
     }
 }
 
-fn choose_existing(node: &Node) -> [Command; 2] {
-    [choose_one_existing(node, 0), choose_one_existing(node, 1)]
+fn choose_existing(node: &Node, config: &ScoreConfig) -> [Command; 2] {
+    [
+        choose_one_existing(node, 0, config),
+        choose_one_existing(node, 1, config),
+    ]
 }
 
-fn choose_one_existing(node: &Node, player_index: usize) -> Command {
+fn choose_one_existing(node: &Node, player_index: usize, config: &ScoreConfig) -> Command {
     let ln_n = (node.score_sum.visit_count as f32).ln();
-    let c = 100.;
     let multiplier = if player_index == 0 { 1. } else { -1. };
-    node.player_score_sums[player_index]
-        .iter()
-        .max_by_key(|(_command, score_sum)| {
-            (multiplier * (score_sum.avg().val + c * (ln_n / score_sum.visit_count as f32).sqrt()))
-                as i32
+    let mut command_confidences =
+        node.player_score_sums[player_index]
+            .iter()
+            .map(|(command, score_sum)| {
+                (
+                    command,
+                    (score_sum.avg() * multiplier).val
+                        + config.explore_exploit_weight
+                            * (ln_n / score_sum.visit_count as f32).sqrt(),
+                )
+            });
+
+    command_confidences
+        .next()
+        .map(|first| {
+            command_confidences
+                .fold(
+                    first,
+                    |(acc_command, acc_confidence), (next_command, next_confidence)| {
+                        if acc_confidence > next_confidence {
+                            (acc_command, acc_confidence)
+                        } else {
+                            (next_command, next_confidence)
+                        }
+                    },
+                )
+                .0
+                .clone()
         })
-        .map(|(command, _score_sum)| *command)
         .unwrap_or_else(|| Command::new(Action::DoNothing))
 }