USV Fleet Planning Considering Logistical Constraints Using Genetic Algorithm

Planning of cooperative missions using a heterogeneous fleet of Unmanned Surface Vehicles (USVs) is complicated by the logistical implications associated with long-duration missions. This paper presents an approach to address these logistical constraints through intelligent planning of at-sea sustainment operations. This is achieved through the use of a low-fidelity predictive digital twin capable of modelling system performance combined with a custom-built genetic algorithm. The proposed system is capable of efficiently generating fleet solutions with adaptability to support a range of scenarios. Several demonstration scenarios are presented with a range of target USVs and support vehicles navigating towards Honolulu Harbor. The results indicate that the planning tool is capable of generating feasible and effective fleet plans that can be implemented by existing USV navigation tools.