PEMODELAN REGRESI NON LINEAR PADA MEMPREDIKSI JUMLAH PENUMPANG KAPAL LAUT DENGAN ALGORITMA GENETIKA

Abstract

Ships are a form of transportation for humans that are used as transportation for archipelagic countries to cross the ocean from one island to another. The number of ship passengers that cannot be estimated every day, creates problems that have a huge impact on shipping management, especially in terms of shortages of vessels and overloading of passengers. For this reason, a model is needed to predict the number of passengers on ships in the future in order to prevent these problems. Therefore, we need a program that can estimate how much the number of ship passengers will increase in the future. In this study, the problem will be modeled using non-linear regression equation modeling consisting of the independent variable (X) and the dependent variable (Y). The regression coefficient is obtained using the concept of a genetic algorithm. In the process using extended intermediate crossover and random mutation, the selection process uses the elitism selection model. The final result is a prediction using Non Linear Regression Modeling with a total RMSE 6.093223783 and MAPE 4917.706862, the best PopSize is 20, the combination of crossover and mutation levels is 0,6: 0,4, and the best number of generations is 150.