我最近进入了神经网络。我目前正在尝试DQN。我已经能够让他们与众多的atari教程一起工作，但发现我无法理解整个问题，因此我为DQN编写了一个简单的模式，该模式仅比B大，如果给出1，如果没有给出1给出0。然后，根据是否正确，将神经网的得分为1或0。不幸的是，我无法让它学习这个简单的问题。有人可以帮助我吗？

# -*- coding: utf-8 -*-
import random
import numpy as np
import env
import gym
from collections import deque
from keras.models import Sequential
from keras.layers import Dense
from keras.optimizers import Adam

EPISODES = 1000


class DQNAgent:
    def __init__(self, state_size, action_size):
        self.state_size = state_size
        self.action_size = action_size
        self.memory = deque(maxlen=2000)
        #self.gamma = 0.95    # discount rate
        self.gamma = 0  # discount rate
        self.epsilon = 0.5  # exploration rate
        self.epsilon_min = 0.01
        self.epsilon_decay = 0.985
        self.learning_rate = 0.001
        self.model = self._build_model()

    def _build_model(self):
        # Neural Net for Deep-Q learning Model
        model = Sequential()
        model.add(Dense(2, input_dim=self.state_size, activation='relu'))
        model.add(Dense(2, activation='relu'))
        model.add(Dense(2, activation='relu'))
        model.add(Dense(self.action_size, activation='relu'))
        model.compile(loss='mse',
                      optimizer=Adam(lr=self.learning_rate))
        return model

    def remember(self, state, action, reward, next_state, done):
        self.memory.append((state, action, reward, next_state, done))

    def act(self, state):
        if np.random.rand() <= self.epsilon:
           return random.randrange(self.action_size)
        action = np.argmax(self.model.predict(state)[0])
        return action

    def replay(self, batch_size):
        minibatch = random.sample(self.memory, batch_size)
        for state, action, reward, next_state, done in minibatch:
            print("stating")
            target = reward
            if not done:
                target = reward + self.gamma * np.amax(self.model.predict(next_state)[0])
            target_f = self.model.predict(state)
            target_f[0][action] = target
            print("Reward: " + str(reward))
            print("Target: " + str(target))
            print(action)
            print(self.gamma*np.amax(self.model.predict(next_state)[0]))
            print(state)
            print(target_f)
            self.model.fit(state,target_f, epochs=1, verbose=0)
        if self.epsilon > self.epsilon_min:
            self.epsilon *= self.epsilon_decay

    def load(self, name):
        self.model.load_weights(name)

    def save(self, name):
        self.model.save_weights(name)


if __name__ == "__main__":
    state_size = 2
    action_size = 2
    #timeRange = 1440
    timeRange = 998
    agent = DQNAgent(state_size, action_size)
    # agent.load("./save/cartpole-master.h5")
    done = False
    batch_size = 500
    totalScore=0
    for e in range(EPISODES):
        env_state = env.GameState()
        env_state.reset()
        state = env_state.step(0)[1]
        #state = np.reshape(state, [1, state_size])
        totalreward = 0
        for time in range(timeRange):
            #if time==timeRange-1:
                #done = True
            action = agent.act(state)
            reward, next_state = env_state.step(action) 

            totalreward += reward
            next_state = np.reshape(next_state, [1, state_size])
            agent.remember(state, action, reward, next_state, done)
            state = next_state
            if done:
                print("episode: {}/{}, score: {}, e: {:.2}"
                      .format(e, EPISODES, env_state.money+env_state.shares*env_state.sharePrice, agent.epsilon))
                totalScore+=env_state.money+env_state.shares*env_state.sharePrice
                done = False
                break
        if len(agent.memory) > batch_size:
            print("replaying")
            print("Total reward: "+str(totalreward))
            totalreward = 0
            #print(agent.memory)
            agent.replay(batch_size)
        # if e % 10 == 0:
        #     agent.save("./save/cartpole.h5")