Sam Brown

HillPark: campus parking that fits your class schedule

INEG 22203, Computing Methods II, Fall 2026
Team project with Ryan Allen and Damir Ashinov

Finding a parking spot at the University of Arkansas depends on your permit, where your classes are, and when you arrive. HillPark is a Java web app that takes your permit and class schedule and, for each class day, ranks the lots you're allowed to use by walking time, driving time, and traffic. It also estimates how full each lot will be when you get there and plans a park-and-ride bus trip from the Economy lot.

Optimization

A p-median integer program in Google OR-Tools picks the best set of lots for the whole week. In the example week, allowing two lots scores 16% better than one.

Machine learning

Gradient boosting and random forest models in Smile forecast traffic slowdowns from logged road speeds, raising R² from 0.32 to 0.51 on held-out days.

Routing

Dijkstra's algorithm on a 13,800-node OpenStreetMap walking network, traffic-aware driving routes, and bus times from the Razorback Transit schedule.

• Java 21
• Vaadin
• Google OR-Tools
• Smile
• Tablesaw
• JUnit
• Git

Golf course resource optimization

Operations Research course project
Scotsdale Golf Course, Bella Vista, AR

Golf courses use a lot of water and bunker sand, and both cost money. This project asks how a course can keep those resources as low as possible while still maximizing its potential profit. I narrowed the scope to a single course, Scotsdale in Bella Vista, so the model can use real conditions instead of generic numbers.

Goal

Minimize water and sand use while maximizing potential profit

Data

Bunker sand quantities estimated from satellite images of the course

Assumptions

The course stays open year-round; a scheduling component is being explored

• Python
• Optimization modeling
• Operations research