Fix error caused by different rlboard.py version in root and solution…

… dir
storagedaemon · Jun 22, 2021 · 40a1686 · 40a1686
1 parent 8ec095c
commit 40a1686
Show file tree

Hide file tree

Showing 4 changed files with 37 additions and 47 deletions.
diff --git a/8-Reinforcement/1-QLearning/README.md b/8-Reinforcement/1-QLearning/README.md
@@ -222,7 +222,7 @@ for epoch in range(5000):
         v = probs(Q[x,y])
         a = random.choices(list(actions),weights=v)[0]
         dpos = actions[a]
-        m.move(dpos)
+        m.move(dpos,check_correctness=False) # we allow player to move outside the board, which terminates episode
         r = reward(m)
         cum_reward += r
         if r==end_reward or cum_reward < -1000:

diff --git a/8-Reinforcement/1-QLearning/rlboard.py b/8-Reinforcement/1-QLearning/rlboard.py
@@ -108,9 +108,10 @@ def is_valid(self,pos):
     def move_pos(self, pos, dpos):
         return (pos[0] + dpos[0], pos[1] + dpos[1])
 
-    def move(self,dpos):
+    def move(self,dpos,check_correctness=True):
         new_pos = self.move_pos(self.human,dpos)
-        self.human = new_pos
+        if self.is_valid(new_pos) or not check_correctness:
+            self.human = new_pos
 
     def random_pos(self):
         x = random.randint(0,self.width-1)

diff --git a/8-Reinforcement/1-QLearning/solution/notebook.ipynb b/8-Reinforcement/1-QLearning/solution/notebook.ipynb
diff --git a/8-Reinforcement/1-QLearning/solution/rlboard.py b/8-Reinforcement/1-QLearning/solution/rlboard.py
@@ -108,9 +108,9 @@ def is_valid(self,pos):
     def move_pos(self, pos, dpos):
         return (pos[0] + dpos[0], pos[1] + dpos[1])
 
-    def move(self,dpos):
+    def move(self,dpos,check_correctness=True):
         new_pos = self.move_pos(self.human,dpos)
-        if self.is_valid(new_pos):
+        if self.is_valid(new_pos) or not check_correctness:
             self.human = new_pos
 
     def random_pos(self):