An open API service indexing awesome lists of open source software.

https://github.com/codelixir/value-iteration

Understanding Markov Decision Processes using Value Iteration algorithm. For MDL course, Spring 2021.
https://github.com/codelixir/value-iteration

Last synced: 3 months ago
JSON representation

Understanding Markov Decision Processes using Value Iteration algorithm. For MDL course, Spring 2021.

Awesome Lists containing this project

README

          

[![forthebadge](https://forthebadge.com/images/badges/ctrl-c-ctrl-v.svg)](https://forthebadge.com)

#### Contributors
[Dhruvee Birla](https://github.com/dhruvxx) and myself.

This assignment was done as a part of the Machine, Data and Learning course, Spring 2021.

---

# Value Iteration

The goal of this exercise is to understand Markov Decision Processes using the Value Iteration algorithm.

The problem statements have been mentioned in [Assignment.pdf](https://github.com/codelixir/value-iteration/blob/main/Assignment.pdf). Task 2 is similar to Task 1, but with the subtasks having a very little variation from the problem in Task 1.

The [notebook](https://github.com/codelixir/value-iteration/blob/main/valueiteration.ipynb) has the code for the main task (Task 1). The scripts for each task are in the [code](https://github.com/codelixir/value-iteration/tree/main/code) directory, with each file having a variation of a small number of lines from each other. The trace files generated from these scripts are stored in [outputs](https://github.com/codelixir/value-iteration/tree/main/outputs). The [report](https://github.com/codelixir/value-iteration/tree/main/report) analyses the algorithm, our approach and the outputs.