Maryo Yoseph Embarto Dwi Osan Sina Pengarang ADRIANA FANGGIDAE Dosen Pembimbing 1 Adriana Fanggidae Ketua Penguji Arfan Yeheskiel Mauko Penguji 2 Bertha Selviana Djahi Penguji 1

The traveling salesman problem (TSP) is one of the optimization problems in everyday life to find the shortest route by taking each city that is passed exactly once. The search for the shortest route can be solved by several algorithms, one of which is the genetic algorithm. Genetic algorithm is an algorithm that applies biological evolution with the aim of getting offspring with better quality than their parents. During the evolutionary process, genetic algorithms are difficult to guarantee that individuals produced from generation to generation are always good. Elitism is used as a way to retain the best quality individuals from generation to generation by duplicating. In this study, the number of duplicates in the best individual in solving the TSP problem was determined. The test is done by comparing the computation time, fitness, and generation of the five elitism conditions used, namely 0%, 10%, 25%, 50%, and 75% of the total population and the parameters used are, jg = 1000, pop = 35, 70, 105, and 140, pc = 0.25, pm = 0.01, threshold = 0.80, n = 30, 50, 80, 120. The results of the system test are 10% of the population of duplicated individuals who are best able to increase the average fitness value by 9% compared to 0% elitism, while 75% elitism provides a faster average computation time of 0.07 seconds in solving TSP problems. Keywords: Traveling salesman problem, genetic algorithm, selection, crossover, mutation, elitism