Dynamic task allocation in multi autonomous underwater vehicle confrontational games with multi-objective evaluation model and particle swarm optimization algorithm

• Published in: Applied soft computing Vol. 153; p. 111295
• Main Authors: Sun, Bing, Zeng, Yuanren, Zhu, Daqi
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2024
• Subjects: Decision making; Game theory; Multi-AUV; Multi-objective evaluation; Particle swarm optimization; Task allocation; Multi-AUV; Multi-objective evaluation; Decision making; Game theory; Particle swarm optimization; Task allocation
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2024.111295

In this paper, a non-zero-sum game model based on a multi-objective evaluation model was adopted to solve the dynamic task allocation problem of underwater multiple Autonomous Underwater Vehicle (multi-AUV). By introducing a functional coordination mechanism, the high degree of coordination in practical confrontation scenarios was reflected, and the autonomous decision-making and task allocation process of multiple AUVs were clarified. Secondly, a multi-objective evaluation model including survival value, strike income, and ammunition loss was constructed, and weighted processing was performed on the multiple objectives using the Analytic Hierarchy Process (AHP) method, obtaining the benefits of each non-zero-sum strategy in dynamic games. In addition, a particle swarm optimization (PSO) algorithm that combines the theory of good point sets theory (G) and speed control factors (S) for improvement, called GSPSO, was used to find the optimal strategy in the game and allocate tasks. Finally, simulation analysis showed that the collaborative system with functional coordination mechanisms significantly improved the combat capabilities of underwater multiple AUVs. The multi-objective evaluation model combined with AHP can correctly and comprehensively evaluate the advantages and disadvantages of each strategy and correctly respond to changing confrontation tasks and decision preferences. The proposed algorithm improved the convergence speed and global search ability by enhancing the diversity of the population and the effectiveness of iterative solutions, ensuring real-time decision-making and task allocation in complex high-dimensional dynamic games. •Game theory method of multi-AUV confrontation in complex underwater environment.•A multi-objective evaluation model to evaluate in a non-zero-sum strategy.•The Analytic Hierarchy Process (AHP) weighting method for decision-making.•The GSPSO algorithm for the Nash equilibrium of the non-zero-sum game.