Stochastic Mixed-integer Programming for Financial Planning Problems using Network Flow Structure

• Main Author: Alimardani, Masoud
• Other Authors: Kwon, Roy H.
• Language: en_ca
• Published: 2014
• Subjects: Financial engineering; Portfolio optimization; 0796
• Online Access: http://hdl.handle.net/1807/44002

Portfolio design is one of the central topics in finance. The original attempt dates back to the mean-variance model developed for a single period portfolio selection. To have a more realistic approach, multi-period selections were developed in order to manage uncertainties associated with the financial markets. This thesis presents a multi-period financial model proposed on the basis of the network flow structure with many planning advantages. This approach comprises two main steps, dynamic portfolio selection, and dynamic portfolio monitoring and rebalancing throughout the investment horizon. To build a realistic yet practical model that can capture the real characteristics of a portfolio a set of proper constraints is designed including restrictions on the size of the portfolio as well as the number of transactions, and consequently the management costs. The model is solved for two-stage financial planning problems to demonstrate the main advantages as well as characteristics of the presented approach.