Weighted strategies to guide a multi-objective evolutionary algorithm for multi-UAV mission planning
Cristian Ramirez-Atencia, Javier Del Ser, David Camacho
TL;DR
This work addresses multi-UAV mission planning by formulating it as a many-objective constraint satisfaction problem (MCMMP) and solving it with a weighted-random MOEA (WRNSGA-II). The method integrates CSP feasibility checks into the MOEA’s fitness evaluation and biases both initialization and mutation through three weight strategies (Arithmetic, Harmonic, Geometric) across three decision facets: the number of UAVs (NUS), distance-based UAV selection (DUS), and distance-based GCS selection (DGS). Encoding uses six alleles to represent task assignments, flight paths, sensors, task orders, GCS mappings, and return-path flight profiles, enabling a rich, constrained search of the solution space. Experiments on 16 realistic datasets show faster convergence and higher hypervolume than naïve NSGA-II, with the best gains from combining geometric NUS and harmonic DUS/DGS, illustrating improved scalability for complex, constrained mission planning and highlighting future work in decision-support for operator selection among Pareto-optimal plans.
Abstract
Management and mission planning over a swarm of unmanned aerial vehicle (UAV) remains to date as a challenging research trend in what regards to this particular type of aircrafts. These vehicles are controlled by a number of ground control station (GCS), from which they are commanded to cooperatively perform different tasks in specific geographic areas of interest. Mathematically the problem of coordinating and assigning tasks to a swarm of UAV can be modeled as a constraint satisfaction problem, whose complexity and multiple conflicting criteria has hitherto motivated the adoption of multi-objective solvers such as multi-objective evolutionary algorithm (MOEA). The encoding approach consists of different alleles representing the decision variables, whereas the fitness function checks that all constraints are fulfilled, minimizing the optimization criteria of the problem. In problems of high complexity involving several tasks, UAV and GCS, where the space of search is huge compared to the space of valid solutions, the convergence rate of the algorithm increases significantly. To overcome this issue, this work proposes a weighted random generator for the creation and mutation of new individuals. The main objective of this work is to reduce the convergence rate of the MOEA solver for multi-UAV mission planning using weighted random strategies that focus the search on potentially better regions of the solution space. Extensive experimental results over a diverse range of scenarios evince the benefits of the proposed approach, which notably improves this convergence rate with respect to a naïve MOEA approach.