MB12* - Defense and Emergency Operations

Overland aircraft Search and rescue (SAR) incidents require rigorous resource allocation. A searchers team deployed to maximize the probability of finding the search object. Using altitude-based deconfliction, multiple SAR units search the same high-priority areas simultaneously, a "stacked search" concept. Several mathematical models are compared with a Benders decomposition algorithm.

Simulation acceleration is essential for optimization over simulation in time-critical contexts such as maritime search and rescue. We explore the acceleration of a search simulation using multithreading and GPU acceleration. Our experiments show solid speedups of up to 260 times compared to a single-threaded baseline.

The design of tactical wireless networks relies on complex interactions between structural constraints, technological choices, and modeling components. Understanding their impact on network performance and topology remains a key challenge.
In this work, we propose a sensitivity analysis under realistic operational conditions. We investigate three dimensions: (i) structural rules, including master hub selection and PMP successor limits; (ii) technological aspects such as antenna beam width and single- versus multi-beam configurations; and (iii) model parameters, including scenario weights and the min–mean trade-off parameter.
Network topologies are generated using a Tabu Search metaheuristic and evaluated through statistical analyses based on Friedman and Wilcoxon tests to assess the significance of variations across different network sizes.
The results reveal scale-dependent effects, threshold behaviors, and distinct roles between topology structuring and performance levels. These findings provide practical insights for model tuning and decision-making in tactical wireless network design.

Defence Research and Development Canada (DRDC) provides scientific and technical advice to the Department of National Defence and the Canadian Armed Forces. In 2025, DRDC developed a method to assess the feasibility of submarine operations in ice-covered waters, supporting quantitative evaluation of design trade‑offs. Using 4,018 days of historical sea-ice data, a region of the Arctic is discretized into polygons classified as open water, marginal ice zone (MIZ), or pack ice. The problem is formulated as a bottleneck path‑finding task under snorting and endurance constraints. Three cases are analyzed: snorting in open water; open‑water snorting with limited under‑ice endurance; and snorting in both open water and MIZ. For each day, breadth‑first search identifies contiguous snortable regions, and a min-max variant of Dijkstra’s algorithm computes the minimum required maximum continuous under‑ice distance. Results show that feasibility is driven primarily by connectivity of snortable regions rather than endurance alone.