Solution method and parameter estimation of uncertain partial differential equation with application to China’s population

• Published in: Fuzzy optimization and decision making Vol. 23; no. 1; pp. 155 - 177
• Main Authors: Yang, Lu, Liu, Yang
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2024; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Calculus of Variations and Optimal Control; Optimization; Hypotheses; Mathematical Logic and Foundations; Mathematics; Mathematics and Statistics; Method of moments; Numerical methods; Operations Research/Decision Theory; Optimization; Parameter estimation; Parameter uncertainty; Partial differential equations; Probability Theory and Stochastic Processes; China; Parameter estimation; Uncertain statistics; Uncertain partial differential equation; Uncertainty theory
• ISSN: 1568-4539; 1573-2908
• DOI: 10.1007/s10700-023-09415-5

Since the concept of uncertain partial differential equations (UPDEs) was proposed, it has been developed significantly and led us to study parameter estimation for UPDEs. This paper proposes a concept of residual of a class of UPDEs, which follows a linear uncertainty distribution. Afterwards, an α -path of a class of UPDEs is introduced and the important result that the inverse uncertainty distribution of solution of a class of UPDEs is just the α -path of the corresponding UPDEs is reached. And a numerical method is designed to obtain the inverse uncertainty distribution of solution of UPDEs. In addition, based on the α -path and the inverse uncertainty distribution, an algorithm is designed for calculating the residuals of UPDEs corresponding to the observed data. Then a method of moments to estimate unknown parameters in UPDEs is provided. Furthermore, uncertain hypothesis test is recast to evaluate whether an uncertain partial differential equation fits the observed data. Finally, the method of moments is applied to modeling China’s population and the fitness of the estimated parameters is verified by using uncertain hypothesis test.