diff --git a/examples/A2C/actor.py b/examples/A2C/actor.py
index cf026747f7eb0d06704332aeb5fc9ef64eb852c9..e85fbc00cc4a0ca962dd4315525e32f99fc68c61 100644
--- a/examples/A2C/actor.py
+++ b/examples/A2C/actor.py
@@ -48,12 +48,7 @@ class Actor(object):
         model = AtariModel(act_dim)
         algorithm = parl.algorithms.A3C(
             model, vf_loss_coeff=config['vf_loss_coeff'])
-        self.agent = AtariAgent(
-            algorithm,
-            obs_shape=self.config['obs_shape'],
-            lr_scheduler=self.config['lr_scheduler'],
-            entropy_coeff_scheduler=self.config['entropy_coeff_scheduler'],
-        )
+        self.agent = AtariAgent(algorithm, config)
 
     def sample(self):
         sample_data = defaultdict(list)
diff --git a/examples/A2C/atari_agent.py b/examples/A2C/atari_agent.py
index 1c0e47772d1c747bcd91dc6bd8e7f9984c0f9010..126f403a1f5163f88564c030c5b442ff3b63eda5 100644
--- a/examples/A2C/atari_agent.py
+++ b/examples/A2C/atari_agent.py
@@ -21,30 +21,22 @@ from parl.utils.scheduler import PiecewiseScheduler, LinearDecayScheduler
 
 
 class AtariAgent(parl.Agent):
-    def __init__(self, algorithm, obs_shape, lr_scheduler,
-                 entropy_coeff_scheduler):
+    def __init__(self, algorithm, config):
         """
 
         Args:
-            algorithm (`parl.Algorithm`): a2c algorithm
-            obs_shape (list/tuple): observation shape of atari environment
-            lr_scheduler (list/tuple): learning rate adjustment schedule: (train_step, learning_rate)
-            entropy_coeff_scheduler (list/tuple): coefficient of policy entropy adjustment schedule: (train_step, coefficient)
+            algorithm (`parl.Algorithm`): algorithm to be used in this agent.
+            config (dict): config file describing the training hyper-parameters(see a2c_config.py)
         """
-        assert isinstance(obs_shape, (list, tuple))
-        assert isinstance(lr_scheduler, (list, tuple))
-        assert isinstance(entropy_coeff_scheduler, (list, tuple))
-        self.obs_shape = obs_shape
-        self.lr_scheduler = lr_scheduler
-        self.entropy_coeff_scheduler = entropy_coeff_scheduler
 
+        self.obs_shape = config['obs_shape']
         super(AtariAgent, self).__init__(algorithm)
 
         self.lr_scheduler = LinearDecayScheduler(config['start_lr'],
                                                  config['max_sample_steps'])
 
         self.entropy_coeff_scheduler = PiecewiseScheduler(
-            self.entropy_coeff_scheduler)
+            config['entropy_coeff_scheduler'])
 
         exec_strategy = fluid.ExecutionStrategy()
         exec_strategy.use_experimental_executor = True
diff --git a/examples/A2C/learner.py b/examples/A2C/learner.py
index b6b034e1059ab3c57e762e63a153087e95ec72db..49a77ddfb4d3da5319417e6349f132c1304ee042 100644
--- a/examples/A2C/learner.py
+++ b/examples/A2C/learner.py
@@ -47,12 +47,7 @@ class Learner(object):
         model = AtariModel(act_dim)
         algorithm = parl.algorithms.A3C(
             model, vf_loss_coeff=config['vf_loss_coeff'])
-        self.agent = AtariAgent(
-            algorithm,
-            obs_shape=self.config['obs_shape'],
-            lr_scheduler=self.config['lr_scheduler'],
-            entropy_coeff_scheduler=self.config['entropy_coeff_scheduler'],
-        )
+        self.agent = AtariAgent(algorithm, config)
 
         if machine_info.is_gpu_available():
             assert get_gpu_count() == 1, 'Only support training in single GPU,\