Shaolong Hu

This study first addresses a capacity expansion problem with surging demand in the context of reoptimizing single-allocation hub-and-spoke networks, in which congestions in both hubs and hub links are incorporated simultaneously. To properly capture the correlation between congestion effects and capacity expansion choices, the Kleinrock average delay function and the Bureau of Public Roads function are employed to explicitly model the congestion cost for hubs and hub links, respectively. Then, the problem is formulated as a non-convex mixed-integer nonlinear program due to multivariable non-convex nonlinear congestion cost functions. To solve the resulting non-convex nonlinear model, seven approximate and exact reformulations based on linear approximations and second-order cone programming (SOCP), are developed by exploiting the structure properties of established model. The devised reformulations can be directly solved by utilizing state-of-the-art solvers. In particular, approximation error guarantees of the linear approximation reformulations that utilize the exponential representation of breakpoints are derived analytically. Finally, the performance of the presented reformulation combinations is demonstrated based on extensive computational experiments using Civil Aeronautics Board and Australia Post data sets. By performing sensitivity analyses of key parameters, the impacts of capacity expansion choices on network reconfigurations are explored, and some managerial insights are extracted.

This paper studies a novel heterogeneous Unmanned Aerial Vehicles (UAVs) team orienteering problem for post-disaster search in flood, integrating environmental risk assessment and precedence constraints. Each affected location requires a sequence of two specialized tasks, namely initial thermal screening followed by visual confirmation, which must be performed in a strict order by a heterogeneous fleet of drones. The rescue profit for each location is defined as a time-decaying function of its comprehensive risk index. We formulate the problem as a mixed-integer programming model and design a tailored Adaptive Large Neighborhood Search (ALNS) to solve this NP-hard problem. A real-world case study in Guangxi, China, and extensive computational experiments demonstrate that this environment-driven approach effectively optimizes drone allocation to maximize lifesaving efficacy. Computational experiments demonstrated the efficiency of the ALNS in obtaining high-quality solutions for large-scale cases. This study provides a decision-support framework that bridging environmental risk analysis with routing optimization.

This paper studies a novel Heterogeneous UAV Team Orienteering Problem for flood disaster search and rescue, integrating environmental risk assessment and precedence constraints. Each affected location may require a sequence of two specialized tasks, namely initial thermal screening followed by visual confirmation, which must be performed in a strict order by a heterogeneous fleet of drones. The rescue profit for each location is defined as a time-decaying function of its comprehensive risk index. We formulate the problem as a mixed-integer programming model and design a tailored Adaptive Large Neighborhood Search (ALNS) heuristic to solve this NP-hard problem. A real-world case study in Guangxi, China, and extensive computational experiments demonstrate that this environment-driven approach effectively optimizes drone allocation to maximize lifesaving efficacy. Computational experiments demonstrated the efficiency of the ALNS in obtaining high-quality solutions for large-scale cases. This study provides a decision-support framework that bridging environmental risk analysis with routing optimization.