Multiple UAVs Forest Fire Reconnaissance Task Assignment Based on Enhanced Adaptive Fireworks Algorithm

The assignment of task for multiple unmanned aerial vehicles (UAVs) during forest fire is a complicated and largescale multi-objective optimization problem, which requires the consideration of multiple factors such as the location of fire points, the scale of fire, the coverage of UAVs and energy consumption. In this paper, multiple UAVs task assignment is transformed into a multi-objective discrete optimization problem and solved with the swarm intelligence optimization algorithm. To amend the weaknesses of low efficiency and easy falling into local optimum of swarm intelligence algorithm, an enhanced adaptive fireworks algorithm (EAFWA) is proposed. The innovations of the proposed algorithm are as follows. First, an adaptive explosion radius transformation strategy is proposed for fireworks with different fitness values. Second, the initialization strategy is optimized, which effectively expand the global search capability. Third, an elite-Gaussian offspring selection method is devised to avoid the algorithm falling into local optimum. Furthermore, a UAVs-enabled mobile edge computing (MEC) system is designed, the objective of which is to maximize coverage while minimizing computing resources and energy consumption. Some well-known algorithm and our proposed algorithm are applied to assignment optimization of Australian fires. The comparison results confirm that EAFWA has great advantages in convergence speed, solution effectiveness and robustness, which can provide insight for assignment optimization problem.