Upload 2 files

env/delivery_env.py

@@ -0,0 +1,107 @@
+import random
+
+class DeliveryEnv:
+
+    def __init__(self):
+        self.state_data = None
+        self.total_reward = 0
+        self.done = False
+        self.current_time = 0
+
+    def _generate_orders(self, n):
+        return [
+            {
+                "id": i,
+                "location": random.randint(1, 10),
+                "prep_time": random.randint(1, 3),
+                "deadline": random.randint(6, 15),
+                "assigned": False
+            }
+            for i in range(1, n + 1)
+        ]
+
+    def _generate_riders(self, n):
+        return [
+            {"id": i, "location": random.randint(0, 5)}
+            for i in range(1, n + 1)
+        ]
+
+    def reset(self, task="easy"):
+        if task == "easy":
+            orders = self._generate_orders(2)
+            riders = self._generate_riders(1)
+        elif task == "medium":
+            orders = self._generate_orders(4)
+            riders = self._generate_riders(2)
+        else:
+            orders = self._generate_orders(6)
+            riders = self._generate_riders(2)
+
+        self.state_data = {
+            "orders": orders,
+            "riders": riders
+        }
+
+        self.total_reward = 0
+        self.done = False
+        self.current_time = 0
+
+        return self.state()
+
+    def state(self):
+        return {
+            "time": self.current_time,
+            "orders": self.state_data["orders"],
+            "riders": self.state_data["riders"]
+        }
+
+    def step(self, action):
+        if self.done:
+            return self.state(), 0.0, True, {"error": "Episode finished"}
+
+        try:
+            parts = action.split("_")
+            order_id = int(parts[1])
+            rider_id = int(parts[3])
+
+            order = next(o for o in self.state_data["orders"] if o["id"] == order_id)
+            rider = next(r for r in self.state_data["riders"] if r["id"] == rider_id)
+
+            if order["assigned"]:
+                return self.state(), -1.0, False, {"error": "Already assigned"}
+
+            # Distance & time
+            distance = abs(rider["location"] - order["location"])
+            delivery_time = order["prep_time"] + distance
+
+            self.current_time += delivery_time
+
+            # 🔥 URGENCY CALCULATION
+            urgency = order["deadline"] - self.current_time
+
+            # 🔥 REWARD COMPONENTS
+            if self.current_time <= order["deadline"]:
+                base_reward = 1.0
+            elif self.current_time <= order["deadline"] + 3:
+                base_reward = 0.3
+            else:
+                base_reward = -0.5
+
+            urgency_bonus = max(0, (10 - urgency) * 0.1)
+            travel_penalty = distance * 0.05
+
+            reward = base_reward + urgency_bonus - travel_penalty
+
+            # Update state
+            order["assigned"] = True
+            rider["location"] = order["location"]
+
+            self.total_reward += reward
+
+            done = all(o["assigned"] for o in self.state_data["orders"])
+            self.done = done
+
+            return self.state(), reward, done, {"error": None}
+
+        except Exception as e:
+            return self.state(), -1.0, False, {"error": str(e)}

graders/grader.py

@@ -0,0 +1,3 @@
+def compute_score(total_reward, max_possible):
+    score = total_reward / max_possible
+    return max(0.0, min(1.0, score))