Loading...
Multi-period Portfolio Optimization Using Model Predictive Control
Jamalzadeh, Saeed | 2018
1231
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 51155 (01)
- University: Sharif University of Technology
- Department: Industrial Engineering
- Advisor(s): Kianfar, Farhad
- Abstract:
- Planning is a process to reach organizational goals, which many organizations can implement it to follow their desired outputs on the considered horizon with optimal costs under existing constraints and minimum control effort. Model Predictive Control (MPC) approach is considered as a planning approach based on forecasting model which contains three parts including objective function, constraints and forecasting model. The aim of using Model Predictive Control is to find specific numbers of current and future inputs on the determined horizon under existing constraints to reach the desired outputs of a dynamic system with minimum control effort. In order to deal with uncertain parameters of the MPC problem, we implemented decision tree under potential scenarios of that parameters. Furthermore, by generating scenarios using t-copula method for state and input matrix of dynamic system in state-space form as parameters in the portfolio optimization problem, we simplified state vector prediction computations through prediction horizon. In addition, we limited volatility of the state of a system in state-space form. All in all, this approach can be applied to many planning problems such as investment management, project control, production planning, and inventory control given a forecasting dynamic model in state-space form. We applied this approach to portfolio optimization problem. Using this approach, our results show that we can track relatively high return on investment by some companies stock trade in NYSE and NASDAQ stock markets under risk constraints
- Keywords:
- Risk Management ; Planning ; Model Predictive Control ; Portfolio Optimization ; Decision Making Tree ; Dynamical Systems ; Dynamic System Analysis
-
محتواي کتاب
- view
- مقدمه
- مرور بر ادبیات
- شرح مدل ریاضی مساله ی کنترل پیش بین مدل
- مساله ی کنترل پیش بین مدل بر پایه ی درخت تصمیم گیری
- مطالعه ی موردی - بهینه سازی پرتفولیو
- نتیجه گیری و پیشنهادات
- منابع و مآخذ
- پیوست کد نرم افزار گمز برای کنترل پیش بین مدل
- پیوست کد نرم افزار گمز برای بهینه سازی پرتفولیو با رویکرد کنترل پیش بین مدل
- پیوست نتایج حل بهینه ی پرتفولیو با استفاده از کنترل پیش بین مدل در گمز
- پیوست داده های مساله ی بهینه سازی سبد سهام
- پیوست کدهای نرم افزار متلب مربوط به پردازش داده های مساله
- پیوست کدهای نرم افزار متلب مربوط به رسم نمودار نتایج مساله