| 刘明，娄德成，王晓飞.基于改进鹦鹉优化算法的船舶推力分配策略研究[J].海洋工程,2026,(1):163~174 |
| 基于改进鹦鹉优化算法的船舶推力分配策略研究 |
| Research on ship thrust allocation strategy based on improved parrot optimization algorithm |
| 投稿时间：2024-11-13 修订日期：2025-03-18 |
| DOI：10.16483/j.issn.1005-9865.2026.01.015 |
| 中文关键词: 推力分配 鹦鹉优化算法 交叉变异 阿基米德螺线 极尽搜索策略 |
| 英文关键词:thrust allocation parrot optimization algorithm（PO） cross-mutation Archimedean spiral exhaustive search strategy |
| 基金项目:国家自然科学基金项目（62273188）；南通市科技计划项目（JC22022085） |
|
| 摘要点击次数: 17 |
| 全文下载次数: 0 |
| 中文摘要: |
| 动力定位系统推力分配求解是一种高度复杂的非线性优化问题，其目标函数和约束条件具有多目标、多约束及非凸特性。传统的推力分配算法在处理该类问题时存在精度低及易陷入局部极值点等问题，而群智能优化算法虽然能够较容易地解决这些问题，但存在收敛速度慢、寻优结果稳定性差和不可靠等问题。针对上述问题，提出一种多策略融合的鹦鹉优化算法（MSPO），该算法通过分段法和改进混沌法相结合初始化种群，不仅增强初始种群的多样性，而且有效保留了种群中的“精英”个体，为算法稳定收敛和可靠收敛奠定基础；对适应度较差的若干个体执行自适应交叉算子策略，有效提升个体寻优效率、加快算法收敛速度；通过随机选取若干个体并采用广域阿基米德螺线更新方式，增强算法在搜索空间中的遍历性，进一步提升算法全局寻优能力；对最优个体实施多尺度多方向的极尽搜索策略，有利于算法在较少迭代次数内获得可靠且稳定的推力分配解。最后以测试函数和Cybership III船模为对象进行改进算法验证，结果表明改进策略提高了算法收敛的可靠性和稳定性，提升了推力分配精度。 |
| 英文摘要: |
| The thrust allocation solution for dynamic positioning systems is a highly complex nonlinear optimization problem, characterized by multi-objective, multi-constrained, and non-convex features in its objective function and constraints. Traditional thrust allocation algorithms exhibit low accuracy and a tendency to fall into local extreme points when dealing with such problems. Although swarm intelligence optimization algorithms can easily solve these problems, they suffer from slow convergence speed, poor stability of optimization results, and unreliability. To address these problems, a multi-strategy fusion parrot optimization algorithm (MSPO) is proposed. The algorithm combines a segmentation method and an improved chaos method to initialize its population, which not only enhances the diversity of the initial population but also effectively retains the “elite” individuals, thereby laying a foundation for the algorithm's stable and reliable convergence. For several individuals with poor fitness, an adaptive crossover operator strategy is adopted to effectively improve optimization efficiency and accelerate the convergence speed. By randomly selecting several individuals and employing a wide-area Archimedean spiral update method, the algorithm's ergodicity in the search space is enhanced, further improving its global optimization capability. Additionally, the algorithm implements a multi-scale, multi-directional exhaustive search strategy for the optimal individual, which facilitates obtaining reliable and stable thrust allocation solutions with fewer iterations. Finally, the improved algorithm is validated using test functions and the Cybership III ship model. The results show that the improved strategy enhances the reliability and stability of algorithm convergence, and improves the precision of thrust allocation. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
