Constraint-Aware Hierarchical Assignment and Routing for Multi-UAV Missions with Time Windows: A Deterministic Simulation Study

Abstract

Joint UAV–target assignment and route construction must account for service windows, payload limits, flight range, and the distance required to return to the depot. We study this problem in a static simulator and evaluate constraint-aware hierarchical assignment and routing (C-HAR), a deterministic constructive heuristic. At each decision step, C-HAR removes UAV–target pairs that fail the current feasibility checks and ranks the remaining pairs using target value, incremental travel, time-window slack, and predicted route imbalance. The evaluation uses 40 shared synthetic instances, independent route replay, Wilson intervals for audited event rates, and paired randomization tests. C-HAR obtained a feasible completed value of 0.941 ± 0.014, with no audited time-window or resource violations among 542 serviced targets. Its feasible-value difference from the equally screened Feasible-Greedy comparator was not statistically significant. The simulator uses fully observed, noise-free states; here, “deterministic” describes the decision rule and state update for a fixed generated instance, not the uncertainty of a physical UAV system. No trained neural policy or flight experiment is evaluated.