This file run Reinforcement Learning (model-free Q-learning) to evaluate the most likely mutation paths from the data

This file run Reinforcement Learning (model-free Q-learning) to evaluate the most likely mutation paths from the data

Usage

mdp_Q_learning_with_linklist(df, discount, N)

Arguments

df

this is the adjacency link link to

discount

the factor determines how much future mutations impact current actions. 0 only looks at next step, close to 1 will have a very long horizon time

N

Number of iterations, also called episodes to explore the MDP

Value

A dataframe that contains the MDP adjacency list, normalized Q-values and mutations taken