A new heuristic based on a parallel implementation of firefly algorithm
MetadataShow full item record
In this chapter, a new optimization algorithm is introduced, inspired by the Firefly Algorithm, which is based on the behavior of fireflies, especially on their flashing characteristics. This new algorithm runs C parallel executions of the algorithm that, after a determined number of iterations, are paused and their results are compared so that, using different restarting techniques, new positions for the fireflies in each thread are determined to continue the executions. This parallel implementation allows us to have more diversity in the experiments, allowing the Firefly Algorithm to obtain better results. For testing the restarting techniques mentioned before, we used 6 multimodal benchmark functions: Step, Sphere, Sum Square, Trid 10, Zakharov, and Rosenbrock functions. For this chapter, different tests were performed to determine the value of the parameters needed to run the algorithm. These values were used to run the final experiments. The results obtained with the implementation of our algorithm with the different restarting techniques and the parameters established are compared to the results published by other authors so that we were able to prove the efficiency of this new algorithm. We obtained significantly better results in some of the test functions, proving that our algorithm is a breakthrough in nature-inspired algorithms. © Springer Nature Switzerland AG 2020.
The following license files are associated with this item:
Mostrando ítems relacionados por Título, autor o materia.
Guerrero Valadez, Juan Manuel; Martinez-Rios, Felix (Springer Science and Business Media Deutschland GmbH, 2020-09-09)This chapter introduces a new implementation of the Rain-Fall Optimization Algorithm (RFO) proposed by Kaboli, Sevbaraj, and Rahim in “Rain-Fall Optimization Algorithm. A Population-Based Algorithm for Solving Constrained ...
Martínez Ríos, Félix Orlando; Murillo Suárez, Alfonso (Springer New York LLC., 2019-05)This paper introduces a faster and more efficient algorithm for solving a two-dimension packing problem. This common optimization problem takes a set of geometrical objects and tries to find the best form of packing them ...
A new heuristic algorithm to solve Circle Packing problem inspired by nanoscale electromagnetic fields and gravitational effects Martinez-Rios, Felix; Marmolejo-Saucedo, José-Antonio; Murillo-Suárez, Alfonso (Institute of Electrical and Electronics Engineers Inc., 2019)In this paper, we present a new algorithm for the fast and efficient solution of the Packing problem in two dimensions. The packing problem consists in finding the best arrangement of objects (many geometrical forms) in a ...