PORTFOLIO OPTIMIZATION OF ASABRI AND JIWASRAYA WITH GOLDEN SINE ALGORITHM

• Main Author: Yosua Siregar, Andro
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/57705
• Institution: Institut Teknologi Bandung

Goverment insurance service companies, ASABRI and Jiwasraya, suffered negative equities due to having investments in penny stocks, which was revealed in early 2020. Penny stocks are the kind of stocks which price changes are manipulated by market practicioners for certain interests, resulting in these type of stocks categorized as unusual market activity by Indonesia Stock Exchange. This final assignment will discuss about finding the optimal return and risk values of ASABRI and Jiwasraya portfolios using the Mean-Variance Model which was developed by Harry Markowitz in 1952. Single-objective optimization are executed using three constraints to find the optimal values. The three constraints are (1)buy-in threshold, used to limit the minimum proportion of each stocks, (2)cardinality, used to choose a certain amount, K out of the total of N stocks to buy, (3)roundlot, used to determine the optimal value of a portfolio investment if stocks are being purchased using lot system. Single-objective optimization is used either to minimize portfolio risk with a fixed return value, or to maximize portfolio return with a fixed risk value. Meanwhile, multi-objective optimization is used to minimize portfolio risk and maximize portfolio return simultaneously with the weighted-sum method. For multi-objective optimization, there are 11 weight values used to build a Pareto front, a graph displaying the relation between sigma and return, in which sigma is the squared root of risk. Pareto front shows that return will increase at the sam time risk increases. In this final assignment, the Indonesia Stock Exchange index (IDX 30) and the LQ45 index will be optimized using multi-objective optimization with cardinality constraint. The amount of stocks chosen is set to ????=9 and ????=14. With this, each portfolios created using multi-objective optimization for the IDX 30 stock data can be compared to the ASABRI and Jiwasraya portfolios. This portfolio optimization problem is solved using Golden Sine algorithm, which is a metaheuristic optimization method that applies the principles of sine wave function with decreasing amplitude for the exploration in the solution domain and the golden section method for the exploitation of areas near the global optimum. The algorithm is evaluated using three benchmark functions, Mixed Integer Non-linear Programming problems, and benchmark portfolio optimization problem using buy-in threshold and roundlot constraints