Research on Complex Financial Decision Making Driven by Geometric Aggregation and Intelligent Optimization of High- Dimensional Expert Information

Abstract In complex fields such as finance, multi-attribute group decision making (MAGDM) often encounters challenges like high-dimensional expert opinions and significant information uncertainty. Traditional linear aggregation methods frequently lead to information loss and distorted results in the presence of expert disagreement and outliers. To address these issues, this paper proposes a novel aggregation framework that integrates spatially optimal set points with a plant growth simulation algorithm (PGSA). Expert opinions, expressed as Picture Fuzzy scores, are mapped into high-dimensional space, and the optimal aggregation node is identified by simulating the plant’s dynamic growth towards light. Using financial decision scenarios, the paper systematically compares the proposed method with mainstream linear operators across multiple public and simulated datasets, evaluating performance through metrics such as Hamming distance, cosine similarity, information energy, and correlation coefficient. Results demonstrate that the proposed method enhances aggregation accuracy, robustness, and interpretability, especially in cases of extreme expert disagreement, providing a valuable tool for group decision making in complex, high-dimensional environments.