Application of Probability-based Multi-objective Optimization in Portfolio Investment and Engineering Management Problems

Abstract

Markowitz's approach could not deal with the overall optimization of both minimizing variance of return rate and maximizing return rate simultaneously due to its short of rational method for handling multi-objective optimization. In this article, a hybrid of the probability-based multi-objective optimization with the uniform design for experiments of mixtures is performed to solve the portfolio investment problem of concurrent optimization of both maximizing return rate and minimizing variance of return rate. The probability-based multi-objective optimization is employed to transfer the bi-objective problem of portfolio investment into a mono-objective one with total preferable probability as the goal for the overall optimization of the system in spirit of probability theory, the uniform design for experiments of mixtures is used to perform the subsequent discretization. Project management problem is rather a multi-objective to conduct naturally. The analysis shows the rationality of the hybrid solution.