Mixture experimental designs have been widely used in industry, agricultureand science test, this paper mainly applies mixture experimental designs theory andmethod to the portfolio investment, which is a new perspective to research on theportfolio investment.This paper includes five chapters: The first chapter introduces the research back-ground, contents and significance.In chapter2, we introduce mixture experimental design theory in detail, includ-ing the common mixture designs, the variance function of predicted valve for regres-sion equation, G-optimal designs.In chapter3, we review the domestic and foreign literature on the modern port-folio theory development and methods, and the minimized risk model of securitiesinvestment.In chapter4, we analyze portfolio investment at homoscedastic and het-eroscedastic mixture model. We give the explanation of portfolio risk and portfolioprofit from the mixture designs theory. According to the theory analysis, the max-imum portfolio risk will be minimum at the G-optimal experiment design, and wecan calculate the investment proportional coefficient when portfolio risk minimumthrough some examples.In chapter5, we give the conclusion of the major results in this paper. |