Quality evaluation of linear inequality constrained estimation by Monte Carlo sampling in parameter space. Issue 378 (4th May 2021)

• Record Type: Journal Article
• Authors: Zhang, Songlin; Li, Jingchang; Zhang, Kun; Han, Jie
• Is Part Of: Survey review. Volume 53:Issue 378(2021)
• Page Start: 216
• Page End: 222
• Publication Date: 2021-05-04
• Subjects: Quality evaluation -- Linear inequality constrained EIV model -- Monte Carlo -- Sampling -- Parameter space; Surveying -- Periodicals; Great Britain -- Surveys -- Periodicals; 526.9
• DOI: 10.1080/00396265.2020.1713612
• Languages: English
• ISSNs: 0039-6265

Abstract

Abstract : Existing approaches that describe the quality of inequality constrained estimates are not sufficient to meet the demands of inequality constrained least-square problems. We reconstruct the existing Monte Carlo method by converting the sampling space from observation space to parameter space, and propose a workflow to improve the computational efficiency. The proposed method is verified by a straight-line fitting example with independent and dependent inequality constraints, and the constraints have different intensities. The results show that the proposed workflow has higher computational efficiency than the existing workflow and that the saved time is almost linearly related to the intensity of the inequality constraints.