바이모달 이산정보에 대한 Akaike 정보척도 기반 신뢰성해석

Abstract

A distribution of response usually depends on the distribution of design variables. When a variable follows a distribution with two different modes, response also follows a distribution with two different modes. In this case, methods of reliability analysis developed recently assume that the distribution functions are continuous with a mode. However In actual problems, Since the information is provided often in a discrete form with two or more modes, it is important to estimate the distributions for discrete information with two or more modes. In this paper, we employ finite mixture model to estimate the response distribution with two different modes, and select the best candidate distribution In AIC-based reliability analysis. Mathematical and engineering examples are illustrated to verify the proposed method.