Modifikasi Algoritma Ant Colony Optimization Dalam Menentukan Rute Pengisian Mesin ATM

Abstract

The TSP problem is known as a non-deterministic polynomial-hard (NP-Hard) problem. In its solution, TSP can be solved using swarm intelligence algorithms such as Artificial Bee Colony (ABC), Partical Swarm Optimization (PSO), dan Ant Colony Optimization (ACO). In this study, TSP settlement was carried out using the ACO algorithm because the amount of data was less than 80 data. In addition, modifications were made to the ACO algorithm with the aim of optimizing the probability in node selection by put in Fuzzy C-Means algorithm into the ACO algorithm. 
 Based on the result application of the Modified Ant Colony Optimization algorithm, the distance covered is 101.712 Km when the parameter optimization has been carried out, with parameter values alpha = 5, beta = 0, rho = 0.3, number of ants = 31, dan maximum iteration = 100. Where each parameter has its own role, such as the Intensity Controlling Constant (alpha) which makes the ants only care about the pheromone intensity value without caring about the distance value between points so that the diversity of the paths found gets smaller when the value alpha gets bigger, Visibility Controlling Constant (beta) which affects the diversity of routes produced by each ant where when beta = 0 then the route chosen by each ant varies and when beta > 0 has the possibility for the route that has been selected to be re-elected by other ants so that the diversity of routes found getting smaller, while for the Ant Track Control Constant (rho) it has an influence in determining the next destination point when the value of gets bigger. In addition, the Modified Ant Colony Optimization algorithm has the advantage of accelerating convergence in finding the shortest route.