Hull form optimization is a typical complex engineering problem. The complex design performance space often results in a low optimization efficiency as an optimal solution cannot be ensured. Presently, several methods, such as efficient optimization algorithms, approximate model technology, and high-performance computing are primarily used to reduce the calculation time. However, these methods cannot satisfy practical application requirements in terms of efficiency and solution accuracy. Thus, we investigated a dynamic space reduction optimization framework (DSROF), in this study, wherein data mining is continuously performed during the optimization process to dynamically reduce the range and number of variables. DSROF enables subsequent optimization only in the range that exhibits a high performance, thereby reducing redundant calculations, improving optimization efficiency, and ensuring a higher degree of accuracy. Furthermore, we applied DSROF to function examples and hull form optimization. The results indicate that the use of the DSROF can reduce the calculation cost in hull form optimization by 23% in comparison with that of the particle swarm optimization algorithm.

船型优化是一个典型的复杂工程问题。复杂的设计性能空间往往导致优化效率低下，因为无法确保最优解。目前，主要采用高效优化算法、近似模型技术、高性能计算等多种方法来减少计算时间。然而，这些方法在效率和求解精度方面都不能满足实际应用要求。因此，我们在本研究中研究了动态空间缩减优化框架 (DSROF)，其中在优化过程中不断进行数据挖掘，以动态减少变量的范围和数量。DSROF 只在表现出高性能的范围内进行后续优化，从而减少冗余计算，提高优化效率，保证更高的精度。此外，我们将 DSROF 应用于函数示例和船型优化。结果表明，与粒子群优化算法相比，使用DSROF可以降低船型优化计算成本23%。