Ali Rouhani | Operations Research Scientist • Data Scientist

Montreal, Canada • ali.rouhani@umontreal.ca

Ali Rouhani

Decision Science (Optimization + Predictive & Prescriptive Analytics) • Logistics • Network Design under Uncertainty

Ph.D. candidate in Operations Research at Université de Montréal (CIRRELT) building data-driven decision systems for logistics, transportation, and network design under uncertainty (capacity, demand, disruptions). I develop scalable optimization algorithms (MIP, stochastic/robust planning, decomposition) and combine them with ML workflows (scenario clustering/reduction, parameter tuning) to enable faster what‑if analysis and more reliable plans. Strong implementation background in Julia/Python/C++ with solver integration, reproducible experimentation, and large-scale benchmarking on HPC platforms.

LinkedIn GitHub Scholar

Focus areas

Optimization: MIP, stochastic/robust planning, decomposition

Decision science: logistics, transportation, network design

Data science: clustering-based scenario reduction, evaluation, tuning

Links

ORCID ResearchGate Google Scholar

Keywords

Benders Lagrangian Scenario reduction HPC benchmarking Python/C++/Julia

About

I build decision-support systems that combine predictive signals with prescriptive optimization. My work focuses on logistics and network design problems where uncertainty (capacity, demand, disruptions) matters—and where scalable algorithms are required to deliver answers fast enough for planning cycles.

Impact highlights

Scalability

Decomposition-based solution strategies (Benders/Lagrangian and related primal/dual approaches) to make large models tractable; validated via benchmark-driven computational studies.

Uncertainty → decisions

Scenario-based planning with clustering-driven scenario reduction to improve runtime while preserving decision-quality signals.

Engineering rigor

Repeatable benchmarking workflows (Git + Linux/HPC scripts + standardized logs) enabling reliable post-analysis into tables and figures.

Skills (proven)

Optimization / OR

Mixed-Integer Programming (MIP) Stochastic Programming Robust / Scenario-based Planning Benders & Dantzig–Wolfe Decomposition Lagrangian Relaxation (Math)heuristics

Data Science

pandas NumPy scikit-learn regression clustering tree-based models evaluation & feature engineering forecasting basics scenario reduction

Programming

Python Julia C++ (STL) SQL Java

Solvers / Systems

Gurobi (C++/Python) CPLEX (C++) OR-Tools Linux/Bash Git HPC experiment workflows Delphi (legacy)

I list skills I can defend with real work (not tutorial-only).

Experience

Ph.D. Researcher / Graduate Researcher (Optimization & Data Science) Jan 2021 – Present

CIRRELT, Université de Montréal • Montreal, Canada

Network design Uncertainty Decomposition HPC benchmarking

Formulate logistics/network design and planning models translating operational constraints (capacities, service feasibility, multi‑commodity flows) into prescriptive decisions (investment + routing/assignment).
Implement decomposition-based algorithms (Benders-style and relaxation strategies) on top of commercial solvers to improve tractability versus monolithic formulations; validate through systematic benchmarking.
Develop scenario clustering/reduction workflows to compress large stochastic scenario sets into representative subsets, improving runtime while preserving decision-quality signals.
Apply parameter tuning and prediction/forecasting-style signals where helpful to improve optimization system performance and stability across instance families.
Run extensive computational experiments on HPC environments using Linux scripting and standardized logging; enable reliable post-analysis into tables/figures.

Research Assistant of Operations Research (M.Sc.) Feb 2018 – Dec 2020

Shahed University • Tehran, Iran

Supply chain design Demand uncertainty Decomposition

Designed and implemented decomposition-based solution methods to solve large-scale supply chain design models beyond off-the-shelf MIP limits.
Investigated how demand uncertainty changes solution structure and robustness, comparing deterministic vs robust/scenario-based formulations.
Quantified uncertainty-driven trade-offs (cost vs service feasibility/coverage) and identified when uncertainty shifts key decisions (facility location, capacity sizing, routing policies).

Software Developer / Systems Development (Business Applications) Jun 2017 – Jun 2018

Afrand Sana’t Company • Tehran, Iran

SQL Server Business systems

Developed a CRM and sales management system (Delphi + SQL Server) to improve client tracking and sales reporting.
Optimized SQL queries and stored procedures to reduce reporting latency and improve responsiveness for day-to-day operations.

Selected projects

Logistics Network Optimization Engine C++, Gurobi/CPLEX API, OpenMP, CMake

Up to 60% runtime reduction (benchmarks) • GitHub

Solves large-scale network design problems under capacity and flow constraints (investment + multi‑commodity routing feasibility).
Solver-integrated implementation with performance-oriented C++ design (parallel evaluation, memory-aware routines, efficient data structures) and scenario reduction to lower stochastic complexity.
Enables cost-effective network planning under uncertainty while maintaining service feasibility and customer satisfaction targets.

Collaborative Distribution Network Design under Disruptions Julia, CPLEX API, OOP

~50% runtime reduction • <0.1% optimality gap (tested instances)

Recommends allocation and contingency plans for collaborative distribution platforms under demand/disruption uncertainty.
Combines clustering-based scenario reduction with a scalable heuristic/optimization workflow to accelerate solves while preserving solution quality.
Supports lower inventory and distribution costs across time periods by choosing effective stocking/distribution locations and re‑planning when conditions change.

Humanitarian Relief Distribution Decision Support Java, Algorithms, OOP

~40% runtime reduction • <1% optimality gap (tested instances)

Generates relief allocation and distribution plans for disaster response settings to improve coverage and timeliness.
Models demand/needs patterns from impacted areas and solves the planning problem using matheuristics tailored for sparse networks.
Helps ensure critical supplies are positioned and delivered efficiently, improving responsiveness to affected communities.

Research

Publications

A Lagrangian Decomposition Algorithm for Robust Green Transportation Location Problem

Rouhani, A.; Bashiri, M.; Sahraeian, R. — International Journal of Engineering TRANSACTIONS A: Basics, 32(1), 85–91 (2019). • publisher

Industry translation: robust transportation-location planning under uncertainty (cost/emissions vs service feasibility).
Methods: Lagrangian decomposition.

Working papers

Benders Decomposition for Two-layer Network Design with Capacity Decisions (Presented at ISMP 2024)

Industry translation: scalable network investment and capacity planning; faster planning cycles via decomposition and valid inequalities.
Methods: Benders decomposition, valid inequalities, matheuristics → C++.

Two-layer Network Design under Capacity Uncertainty

Industry translation: risk-aware capacity decisions via stress-tested network design across scenarios.
Methods: scenario generation, decomposition, branch-and-cut → C++.

Nested Benders Decomposition with Lagrangian Cuts for Two-Stage Stochastic Mixed-Integer Programs

Industry translation: faster, stronger bounds for large stochastic planning models.
Methods: nested Benders, Lagrangian relaxation, branch-and-cut → C++.

Dynamic capacity (re)-allocation in collaborative distribution platforms under disruption uncertainty

Industry translation: operational re-planning under disruptions; dynamic capacity allocation when conditions change.
Methods: multi-stage stochastic programming, dual decomposition, scenario generation → Julia.

Accelerated Benders Decomposition for Stochastic Liner Shipping Fleet Repositioning

Industry translation: asset rebalancing under uncertainty to improve cost and service reliability in maritime logistics.
Methods: two-stage stochastic programming, Benders, valid inequalities → Julia.

For the most complete list, see Google Scholar.

Service & activities

Reviewer

European Journal of Operational Research (EJOR)
Omega
International Journal of Production Research (IJPR)
Operational Research

Presentations

ISMP 2024 (conference presentation)
Data-Driven Optimization (Shahed University, 2019)
Unmanned Aerial Vehicles (Shahed University, 2019)

Education & awards

Education

Ph.D. in Computer Science (Operations Research), Université de Montréal — 2021 – Present (Expected May 2026)
M.Sc. in Industrial Engineering (System Optimization), Shahed University — 2016 – 2019
B.Sc. in Industrial Engineering, University of Qom — 2011 – 2016

Awards

UdeM Exemption Scholarship for International Students (2021)
Best Undergraduate Researcher, University of Qom (2016)
3rd Place, Tehran Mathematics Olympiad (2006)