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

  • 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。

     

    Abstract:
    Objectives In view of the unclear maturity assessment methods and ambiguity in assessment of existing ship production design software, a maturity assessment model for ship production design software is proposed.
    Methods Based on the four stages of ship production design process, including hull, piping, outfitting and coating, the maturity assessment system is constructed and the maturity factors at each level are determined. Combined with Bayesian network and fuzzy best worst method, a completely objective weighting method is proposed to improve the accuracy of dataset. 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 local minimum problem, and comprehensively evaluates the maturity of the software.
    Results The case shows the root mean square error of PSO-BP is reduced by 56.86% compared to BP.
    Conclusions The model accuracy and speed is good enough to meet practical needs, and provide a new approach for software maturity assessment in the shipbuilding industry.

     

/

返回文章
返回