APPROXIMATION OF BANK DEPOSIT AND LOAN DYNAMICS MODEL PARAMETERS WITH SINE COSINE ALGORITHM (SCA)

• Main Author: Batara, Stefan
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/49727
• Institution: Institut Teknologi Bandung

Optimization methods have recently shifted towards population-based metaheuristics that utilize stochastic measures. Mirjalili created the Sine Cosine Algorithm (SCA) and evaluated its performance on unconstrained benchmark problems and optimization of airplane wing design. In this paper, a modified method is constructed to allow SCA to optimize Mixed Integer Non-Linear Programming (MINLP) problems. Then, SCA’s performance post-modification is evaluated on general MINLP problems. The algorithm is also utilized to approximate parameters of a dynamic system modeling deposits and loans of banks in Indonesia, which is classified as a MINLP problem. Analysis is then done on the results from a banking context, as well as a computational context. Results show that SCA can be effectively used to approximate the parameters for the model, by analysis of the objective values and a comparison graph between the model and the data.