PORTFOLIO MULTIOBJECTIVE OPTIMIZATION USING MOTH FLAME OPTIMIZATION METHOD

• Main Author: Moranoud Simbolon (NIM: 10114051), Harnosveric
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/27743
• Institution: Institut Teknologi Bandung

A portfolio is a collection of various types of assets such as stocks, deposits, bonds, and so on. In this final project the used assets are stocks. Portfolio optimization is an effort made by investment managers to get the maximum return with the specified risk limit or to get the minimum risk with given the return target. In this final project, portfolio optimization will be carried out by maximizing return while minimizing risk, which is called portfolio multiobjective optimization. Portfolio multiobjective optimization is applied to the Markowitz portfolio model using the ϵ-constraint method. The ϵ-constraint method is a multiobjective method by making only one objective function (single-objective) and the other objective functions used as constraint functions, by setting a boundary value, called ϵ. In terms of portfolio multiobjective optimization in this final project, the objective function of maximizing the return will be used as a constraint function. Portfolio multiobjective optimization conducted in this final project consists of three portfolio cases, namely a simple portfolio, a portfolio with buy-in threshold constraints and portfolio with cardinality constraints. The optimization method that will be used is the Moth-Flame Optimization (MFO) method. This method will be modified to resolve integer optimization problems or Mixed Integer Non-Linear Programming (MINLP). Then, The MFO method will be used to solve multiobjective optimization problems for non-portfolio and portfolio cases. Furthermore, from the results obtained will also be mapped a curve from a collection of multi-objective optimal solutions called pareto fronts.