Supply Chain Models and Systems

Moving Average Forecasting

This was my first hands-on experience using Google Colab and Python to solve a supply chain problem. I implemented a simple moving average model to smooth out demand fluctuations and understand basic forecasting. This beginner-level exercise helped me get comfortable with Python syntax, logic, and the power of cloud-based data tools to lay the groundwork for more complex modeling later in the course.

Demand Forecasting with Exponential Smoothing

In this project, I implemented single, double, and triple exponential smoothing techniques to forecast product demand over time. I also compared their performance against linear regression using MAD, MSE, and MAPE. This helped me understand the practical use of forecasting in supply chain planning and how to tune smoothing parameters for seasonal and trend-based data.

Facility Location Optimization using Lagrangian Relaxation

I solved the Uncapacitated Facility Location Problem (UFLP) using Lagrangian Relaxation. By relaxing customer assignment constraints and incorporating dual variables, I determined which facilities to open and how to assign customer zones to minimize total cost. This project sharpened my understanding of optimization models, spatial data analysis, and facility network design.

Bass Diffusion Model for New Product Adoption

I applied the Bass Diffusion Model to estimate weekly and cumulative adoption of a new product over time. The project involved calculating peak demand and plotting adoption curves to visualize diffusion dynamics. It introduced me to product life cycle modeling and how adoption parameters (p, q, m) influence forecasting.

Distance Metrics for Location Analysis

This mini project explored the differences between Euclidean, Manhattan, and Haversine distances. I coded each metric from scratch and calculated distances between geographic coordinates. This built foundational knowledge for spatial analysis in logistics and demand mapping.

Facility Location Optimization with Linear Programming

Using linear programming and the PuLP library, I modeled and solved a facility location problem with fixed opening costs and transportation distances. I calculated which distribution centers should be open and how to serve each demand point, ensuring every customer was served efficiently. This exercise deepened my skills in formulating and solving real-world logistics problems.

Box Office Revenue Prediction with Linear Regression

In this project, I built a multiple linear regression model to predict movie box office revenue based on production budget, genre popularity, and cast star power. I trained the model using scikit-learn and interpreted the regression coefficients to understand feature impact. This project demonstrated the power of data analytics in forecasting and strategic decision-making.

EOQ & Inventory Cost Optimization

I developed an Economic Order Quantity (EOQ) model to determine optimal order size, reorder point, and cost breakdown. I also compared EOQ performance with power-of-two policies and considered backordering strategies. This helped me explore inventory planning under uncertainty and cost trade-offs in procurement.