<![CDATA[Sugar demand will increase in line with the increase in population, income, and growth in food and beverage processing industry. Therefore, in order for the sugar production process is always increasing in accordance with needs of the sugar itself, hence need for production planning. Accurate forecasting can help companies in taking decisions to determine the amount of sugar to be produced, the materials needed and determine the price of the goods. One method that can be used to do the prediction algorithm is Extreme Learning Machine. But that method in a selection of input and weight bias is chosen randomly, this can lead to the results obtained in the calculation less maximum. This need for a combination of Particle Swarm Optimization algorithms that can perform optimization the input value weight and bias optimally. This research uses data 45 milled sugar production with 5 features. Based on the research that has been performed, the obtained optimal parameters, namely the number of population size 50, 80% training data comparison (36), the number of hidden neurons 10, weighs of inertia 0.5, and a maximum of iterations 250. The parameter value is obtained from the average MAPE of 0.59%. From the average MAPE results obtained, shows that the addition of the PSO algorithm on ELM can determine the value of the input of weight and optimal bias.]]>
Copyrights © 2018
