How to implement the Expected Return calculation ?

Hi @Saurabh_Patel1,

Can you elaborate a little on the exact use case and provide some context / example?

Not sure if this answers your question, but let me give it a try:

In general you can calculate your expected return R_e by considering probabilities p_i for each return R_i and sum the products up:

R_e= \sum_i R_i p_i.

Dependent on when your returns are expected, sometimes you need to consider the „cost of time“ and think about whether discounting is needed, see also this thread: Intuition behind the discount factor in reinforcement learning - #2 by Christian_Simonis

Best regards

Christian

In lab 1 of reinforcement learning, i didnot understood how is the Expected return for state 2 is 78.48 and 58.92 when the misstep_prob = 0.2, how did the calculation came down to this ?