Bryan Galarza

Bryan Galarza

Postdoctoral Researcher at the University of Antwerp. Operations Research Specialist.

Education

PhD in Operations Research

University of Antwerp, Faculty of Business and Economics · 2023

Doctoral research on the design and optimization of semi-flexible demand-responsive bus services, conducted jointly with KU Leuven and funded by the Research Foundation Flanders (FWO). Supervisors: Prof. Kenneth Sörensen and Prof. Pieter Vansteenwegen.

Thesis: Towards the Goldilocks Zone of demand-responsive bus services. Traditional bus services are rigid, inflexible and waste capacity when demand is low or scattered. Fully flexible on-demand services are efficient but unpredictable, hard to control financially, and cannot serve passengers who do not book in advance. The thesis studies the space in between, and asks what a semi-flexible feeder service takes to design, to operate in real time, and to make a defensible case for, where such a service has mandatory stops guaranteed within a maximum headway, alongside optional stops served only when there is demand nearby.

The work spans an exact mixed-integer programming approach with column generation, large neighborhood search and a new metaheuristic framework, a systematic review of demand-responsive bus systems, and a case study on a real Antwerp corridor. It is published as a thesis abstract in 4OR, and the complete thesis is available here.

MSc in Engineering Sciences: Operations Research and Industrial Engineering

(Burgerlijk Ingenieur)

Ghent University, Faculty of Engineering and Architecture · 2018

A two-year master’s in the Department of Industrial Engineering Systems and Product Design, with a strong methodological core: optimization models, heuristics, decomposition techniques, queueing theory, and manufacturing and service systems simulation.

Thesis: Multi-modal Coordination Schemes for Intelligent Traffic Systems. Traffic signal control is classically optimized for cars alone. This thesis extends single-mode control to a genuinely multi-modal scheme: a network of three intersections and eleven signals modelled on Ghent’s R20 ring road, where congestion is estimated with Monte Carlo estimators of expected queue sizes, and cars are weighted against public transport. Four models were compared (uni-modal, multi-modal with passenger car equivalents, bus-priority, and a bi-objective variant that weights public transport stops), each optimized with simulated annealing and multi-objective simulated annealing. Supervisors: Prof. E.-H. Aghezzaf and Prof. Sidharta Gautama.

BSc in Engineering Sciences: Chemical Engineering and Materials Science

(Burgerlijk Ingenieur)

Ghent University, Faculty of Engineering and Architecture · 2016

Three years in the engineering sciences programme, specialising in chemical technology and materials engineering (track CTMA). The first year was the common part shared across the faculty’s engineering programmes. The specialisation that followed covered chemical process technology, materials science, transport phenomena and thermodynamics. The engineering projects along the way:

  • Cooling a processor with ion wind (2013): designing and building an ion wind generator to cool a CPU, as a first hands-on exercise in thermal engineering and measurement.
  • The secrets of oil, gas and bio-oil (2014): analysing bio-oil from genetically modified poplar (elemental analysis, GC×GC) and bio-ethylene production, a team project on renewable fuels and their chemical precursors.
  • Microstructure of alloys: TRIP steel (2015): a materials practicum using dilatometry, SEM, XRD and hardness testing to read how a transformation-induced plasticity steel’s microstructure governs its mechanical behaviour.

Selected coursework

Master’s level: Operations Research Models and Methods · Heuristics and Search Methods · Industrial Systems Modeling and Optimization (Dantzig–Wolfe decomposition, Benders, Lagrangian relaxation) · Simulation of Manufacturing and Service Systems · Network Modelling and Design · Queueing Theory · Quality Engineering and Industrial Statistics · Manufacturing Planning and Control · Supply Chain Engineering · Estimation and Decision Techniques.

Bachelor’s level: Mathematical Analysis I–III · Linear Algebra and Geometry · Discrete Mathematics · Probability and Statistics · Numerical Mathematics · Signals and Systems · Modelling and Control of Dynamic Systems · Transport Phenomena · Physics I–II · Process Technology · Materials Technology.

Technical toolkit

Gurobi, CPLEX and Hexaly for exact optimization, C++ and Python (NumPy, pandas, SciPy, matplotlib) for models, heuristics and analysis, LaTeX for writing, and OpenStreetMap for road-network instances and travel times. Docker for containarized systems. Git, GitHub and BitBucket for version control. MongoCxx for database management. Heuristics and metaheuristics developed and applied include large neighborhood search, simulated annealing, GRASP, column generation and tabu search.