The goal of this project is to build a reinforcement learner that can transform a given integer into a target integer using a set of actions (e.g. addition, multiplication). With an initial problem of building a reinforcement learner for data wrangling, the project is an initial attempt for processing integers. Multiple rules and trials show difficulties of the model to converge to a solution, mainly due to the wide range of states that the agent can be in.
As next steps, other rule sets should be explored (different rewards and punishment systems) or other methodologies aside from Q-Learning can be used for the project.
The full report is included in the Jupyter notebook. If you have any questions regarding this study, please send me a message via e-mail or LinkedIn.