SARSA Reinforcement Learning

Prerequisites: Q-Learning technique
SARSA algorithm is a slight variation of the popular Q-Learning algorithm. For a learning agent in any Reinforcement Learning algorithm it’s policy can be of two types:- 

1. On Policy: In this, the learning agent learns the value function according to the current action derived from the policy currently being used.
2. Off Policy: In this, the learning agent learns the value function according to the action derived from another policy.

Q-Learning technique is an Off Policy technique and uses the greedy approach to learn the Q-value. SARSA technique, on the other hand, is an On Policy and uses the action performed by the current policy to learn the Q-value.
This difference is visible in the difference of the update statements for each technique:- 

1. Q-Learning:
2. SARSA:

Here, the update equation for SARSA depends on the current state, current action, reward obtained, next state and next action. This observation lead to the naming of the learning technique as SARSA stands for State Action Reward State Action which symbolizes the tuple (s, a, r, s’, a’).
The following Python code demonstrates how to implement the SARSA algorithm using the OpenAI’s gym module to load the environment.
Step 1: Importing the required libraries

Python3

import numpy as np
import gym

Step 2: Building the environment
Here, we will be using the ‘FrozenLake-v0’ environment which is preloaded into gym. You can read about the environment description here.

Python3

#Building the environment
env = gym.make('FrozenLake-v0')

Step 3: Initializing different parameters

Python3

#Defining the different parameters
epsilon = 0.9
total_episodes = 10000
max_steps = 100
alpha = 0.85
gamma = 0.95

#Initializing the Q-matrix
Q = np.zeros((env.observation_space.n, env.action_space.n))

Step 4: Defining utility functions to be used in the learning process

Python3

#Function to choose the next action
def choose_action(state):
    action=0
    if np.random.uniform(0, 1) < epsilon:
        action = env.action_space.sample()
    else:
        action = np.argmax(Q[state, :])
    return action

#Function to learn the Q-value
def update(state, state2, reward, action, action2):
    predict = Q[state, action]
    target = reward + gamma * Q[state2, action2]
    Q[state, action] = Q[state, action] + alpha * (target - predict)

Step 5: Training the learning agent

Python3

#Initializing the reward
reward=0

# Starting the SARSA learning
for episode in range(total_episodes):
    t = 0
    state1 = env.reset()
    action1 = choose_action(state1)

    while t < max_steps:
        #Visualizing the training
        env.render()

        #Getting the next state
        state2, reward, done, info = env.step(action1)

        #Choosing the next action
        action2 = choose_action(state2)

        #Learning the Q-value
        update(state1, state2, reward, action1, action2)

        state1 = state2
        action1 = action2

        #Updating the respective vaLues
        t += 1
        reward += 1

        #If at the end of learning process
        if done:
            break

In the above output, the red mark determines the current position of the agent in the environment while the direction given in brackets gives the direction of movement that the agent will make next. Note that the agent stays at it’s position if goes out of bounds.
Step 6: Evaluating the performance

Python3

#Evaluating the performance
print ("Performance : ", reward/total_episodes)

#Visualizing the Q-matrix
print(Q)