QuantEcon DiscreteDP Solutions

This project contains solutions to several QuantEcon lectures using the DiscreteDP class from the quantecon Python package. DiscreteDP solves finite-state, finite-action Markov decision processes via value iteration, policy iteration, or modified policy iteration, which makes it a natural fit for the dynamic programming problems covered in these lectures.

Lectures covered

• McCall Model — a worker's optimal search/stopping problem: whether to accept or reject a wage offer each period, given a fixed distribution of offers and the option to keep searching.
• Cake Eating — an infinite-horizon consumption/savings problem: how much of a fixed resource ("cake") to consume each period versus save for the future.
• Income Fluctuation Problem — a consumption-savings problem under income uncertainty and a borrowing constraint, where the agent chooses consumption and asset holdings each period.

Each lecture is worked through as a Jupyter notebook in notebooks/, where the problem is formulated as a discrete dynamic program (states, actions, rewards, and transition probabilities) and then solved with DiscreteDP.

Setup

pip install -r requirements.txt
jupyter notebook notebooks/

Project structure

.
├── notebooks/        # Jupyter notebooks, one per lecture/problem
├── requirements.txt  # Python dependencies
└── README.md