Facility Location-Allocation Problem for Emergency Medical Service With Unmanned Aerial Vehicle

This paper models an operation problem of an unmanned aerial vehicle for the emergency medical service (UEMS) system. The model is set up as a location-allocation problem. The coverage distance and capacity of the UEMS facility are modeled as functions of UAVs assigned. The allocation of the demand point is constrained by the variable coverage distance of each facility. The robust optimization approach is used over the cardinality-constrained uncertain demand, which leads to a nonlinear optimization problem. The UEMS location-allocation problem (ULAP) is reformulated to a solvable problem. An extended formulation and corresponding branch-and-price (B & P) algorithm are also proposed, which strengthen the linear programming relaxation bound. The subproblem of the B & P algorithm is defined as a robust disjunctively constrained integer knapsack problem. Two solution approaches of mixed-integer linear programming reformulation and decomposed dynamic programming are designed for the subproblem. To provide time-efficient solutions for large-scale problems, a restricted master heuristic (RMH) is proposed based on the extended formulation. In computational experiments, the B & P algorithm provided a strong lower bound, and the RMH could find an effective feasible solution within an applicable computation time.

Automated summary

This paper models an operational problem of an unmanned aerial vehicle for the emergency medical service (UEMS) system and proposes a robust optimization approach to solve the uncertain demand problem. Computational experiments demonstrate the effectiveness of the proposed methods.