The autonomous swimming of fish in a complex flow environment is a nonlinear and intricate system, which is the focus and challenge in various fields. This study proposed a novel simulation framework for artificial intelligence fish. It employed a high-precision immersed boundary-lattice Boltzmann coupling scheme to simulate the interactions between fish and flow in real time, and utilized the soft actor-critic (SAC) deep reinforcement learning algorithm for fish brain decision-making module, which was further divided into a vision-based directional navigation and a lateral line-based flow perception modules, each matched with its corresponding macro-action space. The flow features were extracted using a deep neural network based on a multi-classification algorithm from the data perceived by the lateral line and were linked to the fish actions. The predation swimming and the various Kármán gait swimming were explored in terms of training, simulation, and generalization. Numerical results demonstrated significant advantages in the convergence speed and training efficiency of the SAC algorithm. Owing to the closed-loop “perceive-feedback-memory” mode, intelligent fish can respond in real-time to changes in flow fields based on reward-driven requirements and experience, and the accumulated experience can be directly utilized in other flow fields, and its adaptability, model training efficiency, and generalization were substantially improved.

The reef ecosystem plays a vital role as a habitat for fish species with limited swimming capabilities, serving not only as a sanctuary and food source but also influencing their behavioral tendencies. Understanding the intricate mechanism through which fish adeptly navigate the moving targets within reef environments within complex water flow, all while evading obstacles and maintaining stable postures, has remained a challenging and prominent subject in the realms of fish behavior, ecology, and biomimetics alike. An integrated simulation framework is used to investigate fish predation problems within intricate environments, combining deep reinforcement learning algorithms (DRL) with high-precision fluid-structure interaction numerical methods-immersed boundary lattice Boltzmann method (lB-LBM). The Soft Actor-Critic (SAC) algorithm is used to improve the intelligent fish’s capacity for random exploration, tackling the multi-objective sparse reward challenge inherent in real-world scenarios. Additionally, a reward shaping method tailored to its action purposes has been developed, capable of capturing outcomes and trend characteristics effectively. The convergence and robustness advantages of the method elucidated in this paper are showcased through two case studies: one addressing fish capturing randomly moving targets in hydrostatic flow field, and the other focusing on fish counter-current foraging in reef environments to capture drifting food. A comprehensive analysis was conducted of the influence and significance of various reward types on the decision-making processes of intelligent fish within intricate environments.

The study of fish swimming behaviours and locomotion mechanisms holds significant scientific and engineering value. With the rapid advancements in artificial intelligence, a new method combining deep reinforcement learning (DRL) with computational fluid dynamics has emerged and been applied to simulate the fish’s adaptive swimming behaviour, where the complex fish behaviour is decoupled to focus on the fish’s response to the hydrodynamic field, and the simulation is driven by reward-based objectives to model the fish’s swimming behaviour. However, the scale of this cross-disciplinary method is directly affected by the efficiency of the DRL model. To promote it to more general application scenarios, there is a pressing need for further research on more efficient and economical network architectures to address the challenge of approximating state-value function in high-dimensional, dynamic, and uncertain environments. Building upon a previously proposed computational platform for the simulation of fish autonomous swimming behaviour, we integrated Kolmogorov–Arnold Networks(KANs) and tested their performance in point-to-point swimming and Kármán gait swimming environments. Experimental results demonstrated that, compared to long short-term memory Networks(LSTMs) and multilayer perceptron networks(MLPs), the introduction of KANs significantly enhanced the perception and decision-making abilities of the intelligent fish in complex fluid environments. With a smaller network scale, in the point-to-point swimming case, KANs effectively approximated the state-value function, achieving average reward improvements of up to 88.0% and 94.1% over MLPs and LSTMs networks, respectively, and increased by 766.7% and 105.6% in the Kármán gait swimming case. Under comparable network sizes, the intelligent fish with KANs exhibited faster learning capabilities and more stable swimming performance in complex fluid settings.

The study of fish swimming behavior and locomotion mechanisms holds substantial scientific and engineering significance. With the rapid progression of artificial intelligence, the integration of artificial intelligence with high-precision numerical simulation methods presents a novel and highly efficient tool for investigating fish behavior. In this paper, we proposed a fish perception model that more closely reflects natural logic. By introducing the individual vision and partially visibility model, a physics-based visual system that mirrored the sensory capabilities of live fish was developed. Furthermore, through the construction of a flow vision using conventional neural networks, we enhanced the intelligent fish's ability to detect unsteady hydrodynamic parameters via numerical lateral line. The validity of the new model was demonstrated through experiments, which the intelligent fish hunts complex moving targets in unsteady flow. Finally, we applied the model to study the refuge/predation behaviors of coral reef fish under varying unsteady flow pressures. The results indicated that swimming capabilities significantly impact fish survival strategies in high flow velocity, highly unsteady hydrodynamic environments, shaping distinct evolutionary decision-making traits. These insights may help to understand the niche competition mechanisms of fish in different flow conditions.

Fish schooling, an adaptive strategy shaped by long-term evolution, is closely associated with core biological needs such as survival and reproduction. This study, having adopted a numerical simulation framework built by combining the deep reinforcement learning algorithm with the immersed boundary-lattice Boltzmann fluid–solid interaction simulation method, focused on investigating the effects of modal transitions on energy efficiency and array dynamics during large-scale turning of fish schools under low hydrodynamic pressure environments. Initially, strategic control was applied to two fish utilizing body–caudal fin propulsion, aiming to explore their schooling modes and the intrinsic energy-saving mechanisms inherent to each mode. Subsequently, the school size was expanded non-proportionally, and the influence of modal transitions on energy consumption was analyzed through a leader-guided turning strategy. Results indicated that under low hydrodynamic pressure, carangiform fish exhibited three modes in their modal composition, namely, the vortex action mode, the lateral force action mode, and the combined action mode. Additionally, with an increased number of schooling individuals, the modal transitions of carangiform fish demonstrated greater arbitrariness and flexibility. Notably, exclusive reliance on hydrodynamic information perception via the lateral line could enhance efficiency in small-scale schools. However, as the school expanded, the complex flow field limited the ability of this approach to significantly improve overall schooling performance; meanwhile, in low hydrodynamic pressure environments, the collective movement of fish schools does not depend solely on hydrodynamic information. These findings provide theoretical guidance and strategic foundations for the design of clustered bionic underwater intelligent robots with different undulatory propulsion modes.

This study extended a hybrid immersed boundary-lattice Boltzmann method (IB-LBM) to analyze the hydrodynamic and thermodynamic behaviors of a fluid heat exchanger (FHE) with an asymptotic tube configuration. The effects of Reynolds number (Re), superheat levels, porosity, and cylinder arrangements on temperature and flow velocity were primarily examined. Several thermodynamic parameters were introduced as metrics to evaluate the thermal efficiency of the FHE. The results revealed that the Re significantly influenced the hydrodynamic and thermodynamic behaviors of the FHE. Pressure drops increased with higher Re, while the rapid decline in average heat flux was attributed to reduced conductive heat transfer caused by decreasing temperature difference. Moreover, superheat levels directly impacted the outlet temperature but primarily affected the frequency of flow velocity fluctuations. The average flow frequency, determined through the fast Fourier analysis, helps avoid resonance with the structure's natural frequency, thereby preventing potential structural failures. Additionally, five different porosity and cylinder configurations were analyzed to evaluate their impact on thermal performance of the FHE. The findings from this research provide valuable insights for enhancing the thermal efficiency of the FHE.