基于t-SNE加权和的高维多目标优化算法

High-dimensional Multi-objective Optimization Algorithm Based on t-SNE Weighted Sum

传统的基于t-SNE的高维多目标优化算法在简化目标集时,虽然可以大大降低算法计算复杂度,但也可能损失目标集中有意义的部分属性,导致算法准确性降低。为此,对冗余目标和初始化种群双方面进行择优保留,提出了一种基于t-SNE加权和的高维多目标优化算法。利用加权和对t-SNE-NSGAⅡ算法处理的冗余目标集进行拟合,保留了部分种群的目标属性,提高了初始种群的质量,提升了算法的准确性,加快了算法收敛速度。实验表明,在目标超过5个时,基于t-SNE加权和的高维多目标优化算法的准确性和收敛性提升明显。当目标为10个时,空间分布度提升了38.7%。

When the traditional high-dimensional multi-objective optimization algorithm based on t-distributed stochastic neighbor embedding(t-SNE-NSGAⅡ)simplifies the target set,although it can greatly reduce the computational complexity of the algorithm,it may also lose some meaningful attributes of the target set,resulting in the accuracy of the algorithm decreased.Therefore,a high-dimensional multi-objective optimization algorithm based on t-SNE weighted sum was proposed to preserve redundant target and initial population.The weighted sum is used to fit the redundant target set processed by t-SNE-NSGAⅡalgorithm,and the target attributes of part of the population are retained,the quality of the initial population is improved,the accuracy of the algorithm is improved,and the convergence speed of the algorithm is accelerated.Experimental results show that when the number of targets is more than 5,the accuracy and convergence of the high-dimensional multi-objective optimization algorithm based on t-SNE weighted sum are improved obviously.When the number of targets was 10,the spatial distribution increased by 38.7%.