Integration of Double Exponential Smoothing Damped Trend with Metaheuristic Methods to Optimize Forecasting Rupiah Exchange Rate against USD during COVID-19 Pandemic

Abstract

COVID-19 pandemic has brought great changes to the stability of the Indonesian state. The disease not only has an impact on public health but also has the effect of weakening the economic sector. One indicator is the weakening of the rupiah exchange rate against the USD. When the pandemic emerged, the rupiah exchange rate started to weaken, which may encourage investors to reduce investment in Indonesia. Therefore, it is necessary to predict the rupiah exchange rate during the COVID-19 pandemic for the coming period. This study applies the Double Exponential Smoothing forecasting method by adding a damped trend factor. The calculation of the parameters of the method becomes the research optimization problem. This optimization problem is then solved using metaheuristic methods, namely Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The performance of the forecasting model is measured based on the magnitude of the forecast error. This study shows that the PSO algorithm is better at obtaining the optimal parameters for predicting the rupiah exchange rate in the coming period compared to GA. The integration error rate of Double Exponential Smoothing damped trend with PSO is 0.70%, while the error rate for the same method with GA is 0.72%. Thus, the integrated performance of double exponential smoothing with metaheuristic optimization is a more excellent method in predicting the rupiah exchange rate against the USD during the period of the Coronavirus outbreak. Furthermore, the addition of a trend dampening factor to the DES method also significantly increases the forecast accuracy.