基于PSO-BP神经网络的船舶生产设计软件成熟度评估方法  

Maturity evaluation method of ship production design software based on PSO-BP neural network

作　　者：王冲[1] 华德睿 WANG Chong;HUA Derui(School of Naval Architecture,Ocean and Energy Power Engineering,Wuhan University of Technology,Wuhan 430063,China)

机构地区：[1]武汉理工大学船海与能源动力工程学院,湖北武汉430063

出　　处：《中国舰船研究》2024年第S02期216-224,共9页Chinese Journal of Ship Research

摘　　要：[目的]针对现有船舶生产设计软件成熟度评估方法尚不明确、评估存在模糊性等问题,提出一种船舶生产设计软件成熟度评估模型。[方法]该模型根据船舶生产设计过程中船体、管系、舾装和涂装4个阶段,构建成熟度评估体系并确定各级成熟因子。结合贝叶斯网络与模糊最优最劣法,提出一种完全客观的赋权方法以提高数据集的准确性。引入粒子群优化(PSO)算法改进反向传播(BP)神经网络,通过PSO对BP神经网络的权值和阈值进行最优化,避免局部最优问题,并对软件的成熟度进行全面评估。[结果]实例分析表明,PSO-BP比BP评价的均方根误差减少了56.86%。[结论]该模型的精度和速度较好,能够满足实际评估需求,为船舶工业软件成熟度评估提供一种新思路。[Objective]This paper proposes a new maturity assessment model for ship production design software in order to address the problem in which the existing methods are unclear and their assessment is ambiguous.[Methods]Based on the four stages of the ship production design process,namely hull,piping,outfitting and coating,a maturity assessment system is constructed and the maturity factors at each level determined.Combined with the Bayesian network(BN)and fuzzy best-worst method(FBWM),a completely objective weighting method is proposed to improve the accuracy of the dataset.A particle swarm optimization(PSO)algorithm is introduced to improve the back propagation(BP)neural network.The PSO optimizes the weights and thresholds of the BP neural network to avoid the local minimum problem and comprehensively evaluate the maturity of the software.[Results]The results show that the root mean square error of PSO-BP is reduced by 56.86%compared to BP.[Conclusion]The accuracy and speed of the proposed model are good enough to meet practical needs,thereby providing a new approach to software maturity assessment in the shipbuilding industry.

关 键 词：船舶生产设计软件 软件能力成熟度模型 贝叶斯网络–模糊最优最劣法 PSO-BP神经网络 

分 类 号：U662.2[交通运输工程—船舶及航道工程]