| Summary: | 碩士 === 國立高雄科技大學 === 金融資訊系 === 107 === The Robo-Advisor provides services for investors or financial management demanders from risk preference exploration, "ptimal fund portfolio proposal, to dynamic fund investment adjustment. This study intended to transform the fund portfolio investment into a strategy portfolio proposal in the Robo-Advisor, therefore strategy investors could filter the strategies based on preferences in the system developed by this study, and constructed an optimized strategy with portfolio theory and method.
This study used the data from a futures companys strategy platform to constructed strategy optimization solutions by using Markowitzs mean-variance portfolio model, which calculated the expected rewards and risks (including the risk of the strategy itself and the risk of mutual risk) based on the strategys return curve. The optimization goal was to minimize the portfolio risk under the limit of reward, with different strategy selection restrictions, contained: (1) position restriction by strategy type, (2) strategy subject restriction, (3) long or short attribute restriction, (4) portfolio position restriction and (5) maximal drawdown restriction. In this way, the optimal investment weights could be obtained by using the Generalizes Reduced Gradient Method (GRG) to deal with Large-scale nonlinear optimization problem, and using the results of the exhaustive method as a comparison.
|
|---|
