Hosang Song

Operations Research researcher with a PhD from the University of Bath, specialising in vehicle routing, transportation and logistics optimisation, mixed-integer programming, and heuristic algorithm development. My doctoral research introduced new covering-based formulations and solution methods for truck-drone routing and close-enough routing problems. I have experience developing MILP and nonconvex MINLP formulations, branch-and-cut algorithms, valid inequalities, and hybrid large neighbourhood search methods. A manuscript co-authored with Professors Gilbert Laporte, Günes Erdoğan, Alistair Brandon-Jones, and Maria Battarra is currently under review at Computers & Operations Research.

2021–2025: 1. Literature Studies: Containing TSP variants, covering problems (CSP, CSP with Nodes and Segments (CSPNS) ), heterogeneous routing problems (flying sidekick TSP, carrier-vehicle TSP), arc routing problems (ARPs), and close-enough problems (CEARP, CETSP)., 2021–2025: 2. Covering Paradigm for Truck-Drone Routing with En-route Drone Operations: Introduced a covering paradigm for truck-drone routing with en-route operations, where drone deliveries are represented implicitly through precomputed coverage parameters rather than explicit routing, assignment, and synchronization decisions. Formulated three hierarchical CSPNS-D variants as MILP models, developed a postprocessing algorithm to reconstruct drone routes, and proposed a hybrid Large Neighborhood Search and Simulated Annealing framework for large-scale instances., 2021–2025: 3. Two Variants of the Covering Salesman Problem with Nodes and Segments for Solving the Close-Enough Traveling Salesman Problem: Introduced two CSPNS variants for solving the Close-Enough Traveling Salesman Problem. Formulated the Geometric-CSPNS as a nonconvex MINLP and proved its equivalence to the CETSP. Developed the Generalized-CSPNS as a MILP-based discretized approach, incorporating valid inequalities and a branch-and-cut framework. Computational results showed that the proposed method outperformed GTSP-based approaches by providing tighter upper and lower bounds and identifying additional optimal CETSP sequences., 2020–2021: Extended research on the Master thesis: Linearized the quadratic terms in the MIQP and re-designed symmetric version of the problem., 2019–2020: Master Dissertation: Multiple-drones-assisted Last-mile Delivery Problem formulated as an asymmetric MIQP. Research also introduced two MILPs as bounds models and local improvements.

• 2013: Mathematics Project Group Leader – introduced high-dimensional Pascal’s triangles; won 1st place departmental award.
• 2013–2015: Dean’s List Scholarship – Yonsei University, for consecutive years of high academic achievement.

"The Covering Salesman Problem with Nodes and Segments Using Drones."

Manuscript in preparation for submission to the European Journal of Operational Research.