An AUVs path planner using genetic algorithms with a deterministic crossover operator

Path planning is an optimization process in which a path between two points is to be found that results in a user-defined optimum satisfaction of a given set of requirements. For small scale path planning, exact algorithms such as linear programming and dynamic programming are usually adopted which are able to give optimum solutions in short time. However, due to their memory intensive nature and computational complexity, exact algorithms are not applicable for medium to large scale path planning. Meta-heuristic algorithms such as evolutionary algorithms can provide sub-optimum solution without the full understanding of the search space and are widely used in large-scaled path planning. However, extra precautions are needed to avoid meta-heuristic algorithms from being trapped in local optimum points. In this paper, a path planner combining genetic algorithms (GA) with dynamic programming (DP) is proposed to solve an autonomous under-water vehicles (AUVs) path planning problem. The proposed path planner inherits the speed of exact algorithms and the scalable nature of meta-heuristic algorithms. Simulation results show that when comparing with conventional GA-based path planners, the proposed path planner can greatly improve the convergence rate and solution quality.