Ziteng WangAssociate Professor · Industrial and Systems Engineering · Northern Illinois University

Research

My research creates operations research (OR) and artificial intelligence (AI) tools for data analytics and analytical decision making in engineering and business applications.

Research themes

132°94°risk scoregait featuresreportwalking videopose estimationgait parametersscreening result two short smartphone videos in, a screening report out
The iGAIT screening pipeline, after the INSAR 2025 poster: video, pose estimation, joint angles, prediction.

Autism screening and diagnosis

Computer vision and machine learning applied to ordinary walking videos, so that gait-based autism screening can reach children whose families have no access to a specialty clinic. The work runs through AI-LEADS with collaborators in speech-language pathology, physical therapy, and computer science, and studies how culture and development shape what the measurements mean.

Selected publications
weather station transshipment recharge base drone leg car leg delivery weather region
Multi-modal medical delivery across weather regions, with a car leg handing off to a drone at a transshipment point and a recharging stop en route, after Bahmani et al. (2026), Transportation Research Part C.

Logistics and transportation

Routing, last-mile delivery, and distribution operations when travel times, demand, and weather are not known in advance. The work builds optimization and learning models that carriers and shippers can use to plan fleets and routes, and evaluates them on field and industry data.

Selected publications
ABC T=0T=1T=2T=3T=4 load 0.50 load 0.50 shared vehicle · load 0.83 carrier 1 carrier 2 consolidated service
Time-expanded service network in which two carriers consolidate freight onto one vehicle, after Zhang et al. (2026), Transportation Research Part E.

Supply chain network design

Where to locate facilities, how to route flows between them, and how independent carriers can share orders and vehicles. Recent work uses robust and data-driven optimization so a network plan holds up under demand and cost uncertainty without becoming too conservative to use.

Selected publications
overshoot cubic L1 spline fit smoothing fit shape-preserving spline fits off-peak peak M1M2M3 scheduling under TOU tariffs
Left: a cubic L1 spline fit preserves a step that a smoothing fit overshoots. Right: jobs scheduled across off-peak and peak electricity prices.

Algorithms and computational optimization

Approximation algorithms, mixed-integer programming formulations, and spline-based approximation for problems that are hard to solve at scale, from parallel machine scheduling to material slitting and shape-preserving data fitting. A related line of work studies constraint qualifications and canonical duality in nonlinear programming.

Selected publications

See all publications, including work outside these themes.

Methodologies

I also work with industry partners to extract business insights from data, optimize strategic planning and operational processes, reduce costs and maximize throughput. I use analytic tools including machine learning, operations research, industrial engineering, statistics and math.

Sponsors

⇧