Single Sensor Scheduling for Multi-Site Surveillance (PREPRINT)

摘要

This paper is concerned with scheduling a single sensor to visit a number of sites with possibly time-variant dynamics. The paper motivates and presents a mathematical model for a sensor scheduling problem arising in the context of military operations research. The contributions made in the paper include the development and comparison of both deterministic and stochastic sensor scheduling methods. Results obtained in the paper show that the proposed heuristic methods can be used in real-time. A computational study is also provided. A deterministic greedy method produces the lowest cost solutions among the heuristics tested. The greedy methods produce the most predictable schedules, while the stochastic methods produce solutions that are less predictable in terms of site revisit times.