Genetic Algorithm-based AUV Mission Optimisation With Energy and Priority Constraints

Traditional methods for inspecting subsea facilities result in high cost, energy waste, and negative environmental effect. To improve such operations, resident underwater vehicles can be employed for inspection and intervention reducing the dependence of often large support vessels. Due to challenging underwater communication conditions, these systems need to be autonomous and should able to plan and re-plan their sequence of actions based on the current circumstances. We propose an optimized allocation and routing planner using a distance constrained vehicle routing solution based on genetic algorithms (GA). Our strategy approximates a minimum-cost walk around all the installations, returning to a seabed charging station while satisfying distance and priority limitations. The presented GA approach is employed to a use case based on the spatial distribution of subsea infrastructure for an existing oil and gas field in Norway. The method introduces a selection technique and priority-based on chromosomal representation. Total distance traveled, fulfilment of distance and priority constraints, and the quantity of sub-routes to visit all installations are used to evaluate the performance of the route planning routine and it is compared to the similar method Simulated Annealing (SA). The outcomes show that the proposed GA strategy outperforms SA for the use case with the AUV on oil and gas field. The proposed method can improve route planning for resident underwater vehicles used for subsea infrastructure inspection and intervention contributing to bringing resident vehicle technology forward which can reduce weather dependence, reduce the environmental effect and costs associated with subsea operations.