Penaksiran Parameter Distribusi Weibull Menggunakan Algoritma Genetika dan Particle Swarm Optimization

Abstract

This study aims toestimate and compare the results of the estimated parameters of the Weibull distributionusing the Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The Weibull distribution used is a two-parameter Weibull distribution and a three-parameterWeibull distribution. In the GA Algorithm, there are evolution operators such as crossover and mutation. Meanwhile, in PSO algorithm doesn’t. The evolution operators that can optimize complex problems and a very wide search space Evaluation of the two method is carried out by observing the difference in the resulting fitness values.Based on the simulation data from the estimator obtained using the R program, it is found that the two-parameter Weibull distribution parameter and a three-parameterWeibull distribution using a AG algorithm are both used in this distribution. This is supported by the small difference in the fitness value of the GA algorithm obtained compared to the PSO algorithm. The sample space also affects the difference in fitness. The large the sample space, the greater the fitness.