About Me

Hello, I am Chengju Liang (梁程炬), a Transportation PhD candidate at Beijing Jiaotong University. I work on operations research, logistics systems, and AI Agent applications. More plainly, I spend a lot of time trying to turn messy logistics problems into models that can be checked, tested, and explained.

I am still very much learning. My code, my models, and sometimes my research taste all need debugging. If you work on logistics optimization, intelligent decision systems, or agentic workflows, I would be happy to learn from you and exchange ideas. Senior researchers are especially welcome; I promise to bring questions, coffee-level curiosity, and a healthy respect for reality.

Research and Application Interests

My current work follows a small loop that I keep returning to:

Scenario. Understand the logistics process before writing equations.
Model. Express the main constraints without pretending the real world is clean.
Algorithm. Test whether the model can actually produce usable decisions.
Agent. Explore how tool use and state tracking can help humans search, compare, and validate options.

It is not a grand theory yet. It is the way I am learning to connect operations research with practical logistics intelligence: vehicle routing, electric logistics fleets, dynamic scheduling, robust optimization, candidate explanation, and result validation.

News

2026: I organized Campus Delivery Agent / 交大微澜 as a public-safe case study of conversational ordering, deterministic logistics decisions, and a runnable local full-stack demo.
2026: I organized Freight Matching Agent as my main public case study for AI Agent × logistics decision support; competition result: Top 24 / 963, finalist award.
2026: I opened my-quant-lab as a research-only Python workbench for backtesting, out-of-sample validation, rolling validation, and human-reviewed AI analysis.
2026: ICTTS 2026 EI paper accepted: Electric Vehicle Pickup and Delivery Routing Optimization Considering Queuing Effects and Wireless Charging Lanes.
2026: ASCE Journal of Urban Planning and Development paper accepted / in production, DOI: 10.1061/JUPDDM/UPENG-6343.
2026: Public-safe GitHub profile and academic homepage were organized for job-search and research communication.

Educations

2024 - 2028, PhD student in Transportation, Beijing Jiaotong University. Research focus: urban smart logistics, vehicle routing, dynamic scheduling, distributionally robust optimization, and AI Agent applications for logistics decision support.
2022 - 2024, Master student in Transportation, Beijing Jiaotong University. Research focus: logistics facility location optimization. Ranked 14 / 152 and received the Graduate National Scholarship.
2018 - 2022, Bachelor student in Transportation, Dalian Maritime University. Focused on routing and transportation planning; recognized as an outstanding graduate.

Selected Publications

(2) Electric Vehicle Pickup and Delivery Routing Optimization Considering Queuing Effects and Wireless Charging Lanes. The 12th International Conference on Traffic & Transportation Studies (ICTTS 2026 EI), accepted.

Keywords: electric vehicle routing, pickup and delivery, queuing effects, wireless charging lanes.
(1) Three-Party Evolutionary Game Analysis on Logistics Sprawl Intervention and Mitigation Framework for Sustainable Urban Distribution. Journal of Urban Planning and Development, ASCE, accepted / in production. DOI: 10.1061/JUPDDM/UPENG-6343.

Keywords: logistics sprawl, sustainable urban distribution, evolutionary game, public-sector logistics governance.

Selected Projects

Campus Delivery Agent / 交大微澜. Led product positioning and full-stack demo delivery for a conversational campus delivery Agent. The local prototype connects structured order drafts and POI validation with price, tidal ETA, routing, and heuristic dispatch. Boundary: six simulated vehicles; no production deployment or real vehicle integration is claimed. [showcase]
Freight Matching Agent. A public-safe case study of an Agentic AI prototype for continuous freight-searching scenarios. The project emphasizes task decomposition, state tracking, tool-use workflow, candidate explanation, and result validation. Evidence: Top 24 / 963, finalist award. [project]
my-quant-lab. A local Python research workbench for A-share and US ETFs, covering data-quality checks, strategy backtests, parameter stability, out-of-sample and rolling validation, candidate comparison, reports, and a Streamlit dashboard. Boundary: research only; no broker connection, automated orders, credentials, or return claims. [project]
Personal Website. A lightweight static GitHub Pages site that maps my research, publications, project evidence, and technical direction. It intentionally contains no resume download, private contact channel, application record, or restricted material. [website]

Honors and Awards

Graduate National Scholarship, 2023.
China International College Students' “Internet+” Innovation and Entrepreneurship Competition, Beijing regional second / third prize, team rank 2, 2022.
National College Student Mathematics Competition, national second prize and provincial awards.
Outstanding Graduate, Dalian Maritime University, 2022.

Skills

Optimization: VRP, EVRPTW, MVRPTW, EMVRPTW, MILP, SAA, Wasserstein DRO, CVaR, robust recoverable scheduling, TOPSIS.
Algorithms and experiments: Python, MATLAB, PyTorch, tabu search, ALNS-style local repair, simulation experiments, GitHub-based project packaging.
AI Agent workflow: task decomposition, prompt engineering, state tracking, tool-use workflow, logistics business rules, and result validation.

Public Boundary

A small boundary note: materials on this homepage are intentionally public-safe. Please do not reuse my private photos, resume files, unpublished documents, enterprise materials, screenshots, credentials, or project details without my permission. I am happy to share more context in a proper interview or collaboration setting, but GitHub is not the place for private data. Forked or external-code repositories are references, not original flagship projects.