mddpg_agent.py

import numpy as np
import random
import os
import torch
import torch.nn.functional as F
import torch.optim as optim
import copy
import torch

from collections import namedtuple, deque
from model_new import Actor_net, Critic_net

class Actor:

    def __init__(self, 
        device,
        agent_ID,
        state_size, action_size, random_seed,
        memory, noise,
        lr, weight_decay):   

        self.DEVICE = device
        self.agent_ID = agent_ID

        self.state_size = state_size        
        self.action_size = action_size
        self.seed = random.seed(random_seed)

        # Hyperparameters
        self.LR = lr        
        self.WEIGHT_DECAY = weight_decay

        

        # Actor Network (w/ Target Network)
        self.local = Actor_net(state_size, action_size, random_seed).to(self.DEVICE)
        self.target = Actor_net(state_size, action_size, random_seed).to(self.DEVICE)
        self.optimizer = optim.Adam(self.local.parameters(), lr=self.LR)


        # Replay memory        
        self.memory = memory

        # Noise process
        self.noise = noise

    def act(self, state, add_noise=True):
        """Returns actions for given state as per current policy."""
        state = torch.from_numpy(state).float().to(self.DEVICE)

        self.local.eval()
        with torch.no_grad():
            action = self.local(state).cpu().data.numpy()
        self.local.train()        

        if add_noise:
            action += self.noise.sample()
    
        return np.clip(action, -1, 1)

    def step(self, state, action, reward, next_state, done):
        """Save experience in replay memory, and use random sample from buffer to learn."""
        # Save experience / reward
        self.memory.add(state, action, reward, next_state, done)

    def reset(self):
        self.noise.reset()
     


class Critic:        

    def __init__(self, 
        device,
        state_size, action_size, random_seed,        
        gamma, TAU, lr, weight_decay):        

        self.DEVICE = device

        self.state_size = state_size        
        self.action_size = action_size
        self.seed = random.seed(random_seed)

        # Hyperparameters
        self.GAMMA = gamma
        self.TAU = TAU
        self.LR = lr
        self.WEIGHT_DECAY = weight_decay

        
        # Critic Network (w/ Target Network)
        self.local = Critic_net(state_size, action_size, random_seed).to(self.DEVICE)
        self.target = Critic_net(state_size, action_size, random_seed).to(self.DEVICE)
        self.optimizer = optim.Adam(self.local.parameters(), lr=self.LR, weight_decay=self.WEIGHT_DECAY)


    def step(self, actor, memory):
        # Learn, if enough samples are available in memory  
        experiences = memory.sample()        
        if not experiences:
            return

        self.learn(actor, experiences)

    def learn(self, actor, experiences):
        """Update policy and value parameters using given batch of experience tuples.
        Q_targets = r + γ * target(next_state, actor_target(next_state))
        where:
            actor_target(state) -> action
            target(state, action) -> Q-value
        Params
        ======
            experiences (Tuple[torch.Tensor]): tuple of (s, a, r, s', done) tuples 
            gamma (float): discount factor
        """
  
        states, actions, rewards, next_states, dones = experiences

        # ---------------------------- update critic ---------------------------- #
        # Get predicted next-state actions and Q values from target models
        actions_next = actor.target(next_states)
        Q_targets_next = self.target(next_states, actions_next)
        # Compute Q targets for current states
        Q_targets = rewards + (self.GAMMA * Q_targets_next * (1 - dones)) #more accurate as it uses the current reward, that is already known and adds the predicted Q-value for the next state action pair
        # Compute critic loss
        Q_expected = self.local(states, actions)#predicted Q-value (state-action value of next state with predicted actions for next state). this ONLY relies on the prediction. No element that is known
        critic_loss = F.mse_loss(Q_expected, Q_targets) #loss function: deviation of prediction from more accurate prediction(Q_targets)
        # Minimize the loss
        self.optimizer.zero_grad()#setting gradient to zero
        critic_loss.backward()#compute gradient (all partial derivatives)
        torch.nn.utils.clip_grad_norm(self.local.parameters(), 1) #gradient clipping
        self.optimizer.step() #adjust weights with the clipped gradient in order to perform gradient descent

        # ---------------------------- update actor ---------------------------- #
        # Compute actor loss
        actions_pred = actor.local(states)
        actor_loss = - self.local(states, actions_pred).mean() #loss function = negative reward --> minimizing leads to maximizing reward
        # Minimize the loss
        actor.optimizer.zero_grad()#same as in critic
        actor_loss.backward()
        actor.optimizer.step()

        # ----------------------- soft-update target networks ----------------------- #
        self.soft_update(self.local, self.target) #self = critic
        self.soft_update(actor.local, actor.target)

    def soft_update(self, local_model, target_model):
        """Soft update model parameters.
        θ_target = τ*θ_local + (1 - τ)*θ_target
        Params
        ======
            local_model: PyTorch model (weights will be copied from)
            target_model: PyTorch model (weights will be copied to)
            tau (float): interpolation parameter 
        """
        tau = self.TAU
        for target_param, local_param in zip(target_model.parameters(), local_model.parameters()):
            target_param.data.copy_(tau*local_param.data + (1.0-tau)*target_param.data)



class OUNoise:
    """Ornstein-Uhlenbeck process."""

    def __init__(self, size, seed, mu=0., theta=0.15, sigma=0.2):
        """Initialize parameters and noise process."""
        self.size = size        
        self.mu = mu * np.ones(size)
        self.theta = theta
        self.sigma = sigma        
        self.seed = random.seed(seed)
        
        self.reset()

    def reset(self):
        """Reset the internal state (= noise) to mean (mu)."""
        self.state = copy.copy(self.mu)

    def sample(self):
        """Update internal state and return it as a noise sample."""
        x = self.state        
        dx = self.theta * (self.mu - x) + self.sigma * np.random.standard_normal(self.size)
        self.state = x + dx
        
        return self.state
        


class ReplayBuffer:
    """Fixed-size buffer to store experience tuples."""

    def __init__(self, device, action_size, buffer_size, batch_size, seed):
        """Initialize a ReplayBuffer object.
        Params
        ======
            buffer_size (int): maximum size of buffer
            batch_size (int): size of each training batch
        """
        self.DEVICE = device

        self.action_size = action_size
        self.memory = deque(maxlen=buffer_size)  # internal memory (deque)
        self.batch_size = batch_size
        self.experience = namedtuple("Experience", field_names=["state", "action", "reward", "next_state", "done"])
        self.seed = random.seed(seed)
    
    def add(self, state, action, reward, next_state, done):
        """Add a new experience to memory."""        
        e = self.experience(state, action, reward, next_state, done)
        self.memory.append(e)
    
    def sample(self):
        """Randomly sample a batch of experiences from memory."""
        if len(self.memory) <= self.batch_size:
            return None

        experiences = random.sample(self.memory, k=self.batch_size)

        states = torch.from_numpy(np.vstack([e.state for e in experiences if e is not None])).float().to(self.DEVICE)
        actions = torch.from_numpy(np.vstack([e.action for e in experiences if e is not None])).float().to(self.DEVICE)
        rewards = torch.from_numpy(np.vstack([e.reward for e in experiences if e is not None])).float().to(self.DEVICE)
        next_states = torch.from_numpy(np.vstack([e.next_state for e in experiences if e is not None])).float().to(self.DEVICE)
        dones = torch.from_numpy(np.vstack([e.done for e in experiences if e is not None]).astype(np.uint8)).float().to(self.DEVICE)

        return (states, actions, rewards, next_states, dones)

    def __len__(self):
        """Return the current size of internal memory."""
        return len(self.memory)