基于改进多表达式编程算法的木材染色配方预测
本研究通过使用计算机配色技术实现对普通木材的高精度染色,从而制造出视觉上与珍贵材无法区分的替代品,以降低对珍贵木材的依赖。在基因表达式编程的基础上,提出了多表达式编程(MEP)算法来预测染料配比,以优化染色后木板的光谱反射率。MEP 算法能够处理多染料之间复杂的交互关系,得到较为直观的函数关系式,对遗传算子概率进行自适应改进,并采用并行程序设计提高函数挖掘效率,在配色预测中得到结果为 0.313的配方相对偏差结果,此外,我们将MEP算法的预测结果与传统神经网络和遗传算法进行了比较,结果表明MEP算法具有优异的预测性能。
This study aims to achieve high-precision coloring of ordinary wood using computer coloration technology,creating substitutes that are visually indistinguishable from precious wood,in order to reduce dependence on precious wood.Based on gene expression programming,a multi-expression programming(MEP)algorithm is proposed to predict dye formulations and optimize the spectral reflectance of dyed wood boards.The MEP algorithm can handle complex interactions among multiple dyes and obtain more intuitive functional relationships.It improves the adaptive probability of genetic operators and employs parallel program design to enhance function mining efficiency.The predicted relative deviation of the formulation in color prediction using MEP algorithm is 0.313,which outperforms traditional neural networks and genetic algorithms.
评论