Delaunay triangle division and TOPSIS-based sequential sampling method

Abstract

The invention discloses a Delaunay triangle division and TOPSIS (technique for order preference by similarity to ideal solution)-based sequential sampling method. The method comprises the steps of firstly performing Delaunay triangle division on an initial sample point; obtaining areas of triangles and pre-estimated errors at the centroids; constructing an initial decision matrix according to attributes of the triangles; performing normalization processing on the initial decision matrix to obtain a normalized matrix; obtaining weights of attribute values through an entropy weight method; constructing a weighted matrix according to the normalized matrix; calculating a positive ideal solution and a negative ideal solution according to the weighted matrix: obtaining Euclidean distances from the attribute values of the triangles to the positive ideal solution and the negative ideal solution according to the positive ideal solution and the negative ideal solution; performing calculation according to the Euclidean distances to obtain comprehensive evaluation coefficients of the triangles; taking the centroid of the triangle with a maximum value of the comprehensive evaluation coefficient as a new sample point; and repeating the steps until a total sample quantity reaches a preset target value. According to the sequential sampling method provided by the invention, global exploration and local optimization are weighed, so that the obtained sample point is optimal.

Description

technical field [0001] The invention belongs to the technical field of computer simulation, and more specifically relates to a sequential sampling method based on Delaunay triangle division and ordering of approximate ideal solutions. Background technique [0002] For complex system design, computer simulation is usually used to replace calculation-intensive and physical experiments, but it is time-consuming to use computer simulation for a complex system with multiple inputs and multiple outputs; in order to solve this problem, a proxy model is used to replace the real model , the original system can be simulated at a greatly reduced cost; the quality of the surrogate model has a great influence on the computational cost and convergence characteristics of design optimization based on the surrogate model, and the quality of the surrogate model depends largely on the use of computer The sample points generated by the simulation, therefore, the sampling method plays a key role...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the above defects or improvement needs of the prior art, the present invention provides a sequential sampling method based on Delaunay triangulation and approximation to ideal solution sorting, the purpose of which is to solve the inability of the existing sequential sampling method to solve the global and local problems. Balancing Technical Issues

Images