Optimization of Single Exponential Smoothing using Particle Swarm Optimization and Modified Particle Swarm Optimization in Sales Forecast

The selection of smoothing constant value, especially in the Single Exponential Smoothing (SES) algorithm, is a crucial path in increasing the performance of a forecasting system. Whereas in the SES method, the constant smoothing value is chosen randomly. Some optimization algorithms can be used to determine the best smoothing constant value. One popular algorithm is Particle Swarm Optimization (PSO), which has proven to be an approach that can solve optimization problems well. However, a PSO development method, namely the Modified Particle Swarm Optimization (MPSO) Algorithm, is claimed to have better performance, even producing a 99% reduction in mean square error over the conventional PSO technique in electrocardiogram (ECG) Signals case. To find a better optimization method, this research was conducted to compare the performance of the PSO and MPSO algorithms in producing the best smoothing constant value in the case of sales forecasting. Finally, based on the test result, the MPSO algorithm makes a better MAPE of 17.95 compared to the PSO algorithm, which produces a MAPE value of 19,515 in the built system of forecasting.