Implemantation of firefly algorithm on Arduino Uno

Abstract

Not only getting the optimal solution of a problem, embedding the algorithm on the microcontroller is also expected to work optimally without burdening the system and fast response. Getting a microcontroller specification that matches the complexity of an algorithm is necessary so that the system can execute the algorithm perfectly. Values for the basic parameters of optimization algorithms inspired by nature such as the firefly algorithm (FFA) which are interpreted into variables greatly affect the performance of the microcontroller in obtaining the expected optimal solution. The observed performance of the Arduino Uno microcontroller in running the FFA includes execution time and memory capacity required to obtain optimal values based on changes in absorption coefficient, random parameters, iterations, and population. Changes in the absorption coefficient and random parameters affect the optimal value but do not significantly affect the execution time and memory capacity of Arduino Uno. Iteration changes greatly affect execution time and population changes most affect the performance of Arduino Uno. With a dynamic memory capacity of 2 Kb, the FFA can be run with a maximum range of 50 populations and up to 20 iterations.