Liang Hu

Apply operations research and optimization to product R&D for intelligent transportation and smart cities and cloud computing for Huawei Cloud and 2012 Labs.

Managed data of traffic sensors on Iowa highway.

—Proficiently used Python/R/ArcGIS/Tableau to visualize/clean/analyze transportation big data, e.g., taxi GPS data, highway traffic sensor data, map data, vehicle sensor data. —Future mobility research: Designed agent-based simulation for electric self-driving taxis; Built ILP model to optimize taxi assignment with system efficiency improved by 16%, in Gurobi; leveraged neural networks to learn optimal assignment solutions with solution time shortened by over 90%. —EV research: Implemented queuing theory and MILP model to optimize location and size of EV charging stations with charger utilization increased by 26%; Mined driving patterns from NYC taxi GPS data, trained logistic regression model to predict EV taxi feasibility with 82% accuracy. —Leveraged deep neural networks (CNN) for Iowa highway crash prediction using Google Waze map/traffic sensors/weather data, with 98.6% accuracy. —Proposed multivariate regression model for EV energy consumption using CAN bus data from 18 connected vehicles; developed adaptive cruise control algorithms for self-driving vehicles with energy consumption saved by 10%.

—Intern at Operations Research Lab, GM Global Research & Development. —Car-sharing product improvement: Analyzed spatial-temporal patterns from data of car-sharing orders; Optimized order fulfillment with 11% increase using ILP model and CPLEX. —Future mobility product development: Built ILP model and designed simulation framework for electric self-driving cars in car-sharing/ride-sharing services (assignment, charging, relocation); Greatly contributed the algorithms to product software in Python.

—Researched into transportation electrification at National Transportation Research Center. —Analyzed travel patterns of EV customers from 2017 National Household Travel Survey dataset using R/Spark/SQL; Proposed innovative choice model for charging behavior based on cumulative prospect theory.

Designed and operated a campus car-sharing system with 7 electric cars; Collected and analyzed car-sharing operation data.