THE APPLICATION OF SIMULATED ANNEALING METHOD IN OBTAINING THE LOWEST LOSS POTENTIAL OF STOCK PORTOFOLIO BY MINIMIZING THE VALUE AT RISK

• Main Author: Wijaya, Christoper
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/49045
• Institution: Institut Teknologi Bandung

The problem of optimization in the topic of econophysics in reviewing stock portfolios is to get the highest profit. Optimization to get the highest profit can be viewed in different ways, including focusing on increasing potential profits or focusing on reducing potential losses or Value at Risk (VaR). To reduce VaR, one of the optimization methods used is the Simulated Annealing method which is related to the physical system of metal quality optimization by heating the metal and slowly lowering the temperatur of the system which is done repeatedly, to obtain optimum metal quality. The same principle is used to get the optimum/global minimum VaR value. VaR (potential loss) can be calculated based on the arrangement of the portfolio and the random distribution of weights. By using the Boltzman probability, the VaR can reach the global minimum because at the beginning of the iteration, this probability give programs to explore more looking for the minimum solution, so that it is not trapped in the local minimum. As time goes by, the program is increasingly focused on finding solutions rather than doing exploration, which results that in the end, the VaR generated will converge at one minimum global value. Compared to other methods, the Simulated Annealing method is quite good in suppressing the value of VaR to be quite small at 1,796% and is able to predict maximum potential losses well when using new stock data. The desired portfolio arrangement and weight distribution can be retrieved back through the best VaR value.