A Multi-Population Co-Evolutionary Task Assignment Method Based on Clustering Information Guidance

Multi-UUV task assignment is a difficult optimization problem for multi-variable and multi-objective NP (Non-deterministic Polynomial), and the traditional genetic algorithm has the problems of premature convergence and low efficiency. According to the above problems, a clustering information, multi-group, co-evolution algorithm is proposed for multi-base, multi-target, multi-UUV cooperative reconnaissance task assignment. The algorithm uses k-means clustering results as inspired information to initiate the population, improving the quality of the initial population and the convergence efficiency of the algorithm; by introducing multiple groups to search the solution space, the global optimization and local search ability can be better balanced, and the performance bottleneck of classical genetic algorithm is optimized only by single population. Simulation experiment results show that the algorithm is robust and has high optimization efficiency and can obtain high-quality collaborative task assignment scheme.