基于决策偏好信息的n维Pareto前沿可视化

New graphical visualization of n-dimensional Pareto front based on decision preference information

针对现有可视化技术不能有效显示高维多目标优化问题的难题,提出了一种基于决策偏好信息的n维图表可视化技术。首先,针对两种不同的决策情况分别提出对应的目标信息共享机制;然后,根据目标共享信息将分层后的Pareto前沿,通过子图表进行绘制,有效显示了各维目标数据信息、性能优劣变化趋势和决策者的偏好信息。基于此思想设计了可视化模型并通过实验加以分析,达到了方便决策者对Pareto最优解集进行分析和决策的目的。

In order to overcome the shortcomings that currently visualization technology cannot display high dimension multi-objective optimization problems effectively, a new graphical visualization technology of n-dimensional-Pareto front based on the weighted preference information is proposed. Firstly, the corresponding objective information sharing mechanism for the two different decision-making situations is proposed. Secondly, according to the sharing information the Pareto front will be stratified and drawn by graphs. It effectively shows the data information of each dimension objective, the variation tendency towards performance and the preference information for decision-makers. This paper devises the visualization model and provides the extensive experimental analyses so as to analyze the Pareto optimal solutions for decision-makers.