Fast fuzzy reasoning and optimization of ship conceptual scheme based on stack-like structure

The design of conceptual schemes for ship hull forms often faces challenges in managing the complex coupling relationships among various design parameters. Traditionally, conceptual schemes for different ship hull forms and corresponding requirements rely heavily on the domain knowledge and engineering experience of ship experts. This approach leads to long design cycles and high costs. Therefore, this study proposes a fuzzy reasoning strategy based on a stack-like structure, aimed at rapidly generating initial solutions for ship conceptual scheme.

This strategy employs fuzzy reasoning based on the user's requirements to derive the performance parameter values of the ship hull that meet the user's requirements. These values are then matched with the design scheme library of the parent ship to obtain the conceptual scheme of the ship hull that meets the user's requirements. Drawing upon the principle of stacked generalization, the method introduces a stack-like structure to systematically combine the initially inferred valid information with the input user requirements in a hierarchical manner. This approach reduces reasoning bias by applying multi-stage verification. Finally, a clustering algorithm is designed to optimize the fusion of the performance parameter library and the parent ship design scheme library. This optimization enhances the rationality of the ship conceptual scheme by establishing quantitative relationships between new designs and proven reference solutions.

The numerical examples demonstrate that both the performance parameters of the ship hull derived through reasoning and the corresponding conceptual schemes effectively address the real needs of users.

To address the significant inference bias in performance parameters observed in existing methods, this study proposes a rapid fuzzy inference system based on a stack-like structure for ship conceptual design. This system enhances computational efficiency while effectively reducing reasoning deviations during the inference process. By leveraging hierarchical reasoning layers, it improves decision-making accuracy in the early stages of ship design. Existing methods for ship conceptual schemes design often result in unreasonable coupling relationships among the solutions generated through reasoning technology. This study thoroughly explores the commonalities between the fuzzy reasoning results and the parent ship, ensuring that the user's requirements are met while eliminating the imbalance in the design parameter coupling of the ship conceptual schemes. As a result, the efficiency and rationality of generating ship conceptual scheme solutions are significantly improved.