Search

In this paper, we show a survey of meta-heuristics algorithms based on grouping of animals by social behavior for the Traveling Salesman Problem, and propose a new classification of meta-heuristics algorithms (not based on swarm intelligence theory) based on grouping of animals: swarm algorithms, schools algorithms, flocks algorithms and herds algorithms: a) The swarm algorithms (inspired by the insect swarms and zooplankton swarms): Ant Colony Optimization algorithm - ACO (inspired by the research on the behavior of ant colonies), Firefly Algorithm (based on fireflies), Marriage in Honey Bees Optimization Algorithm - MBO algorithm (inspired by the Honey Bee), Wasp Swarm Algorithm (inspired on the Parasitic wasps), Termite Algorithm (inspired by the termites), Mosquito swarms Algorithm - MSA (inspired by mosquito swarms), zooplankton swarms Algorithm - ZSA (inspired by the Zooplankton) and Bumblebees Swarms Algorithm -BSA (inspired by Bumblebees). b) The school algorithms (inspired by the fish schools): The Particle Swarm Optimization algorithm - PSO (inspired by social behavior and movement dynamics of fish or schooling). c) The flock algorithms (inspired by the bird flocks): the flocking algorithm, and the Particle Swarm Optimization algorithm (inspired on the dynamics of the birds), d) The herd and pack Algorithms (inspired by the mammal herds and packs): bat algorithm (inspired by bat), wolf pack search algorithm - WPS (inspired by wolfs), Rats herds algorithm - RATHA (inspired by Rats), Dolphins Herds Algorithm - DHA (inspired by Dolphins) and the feral-dogs herd algorithm - FDHA (inspired by feral-dogs herd). [ABSTRACT FROM AUTHOR]

In this paper, we propose the application of discriminant analysis to select appropriately the algorithm that better solves an instance of the Job Shop Scheduling Problem. The discriminant analysis was used as a method of machine learning to find the relation between the characteristics of the problem (complexity indicators) and the performance of algorithms. The prediction of the classification obtained of the discriminant analysis was 60%. [ABSTRACT FROM AUTHOR]

The theory of the computational complexity is a subject of international interest of scientific, technological and enterprise organizations. The computational complexity contains diverse elements such as the classes of problems complexity (P, NP, NP-hard and NP-complete, and others), complexity of algorithms (it is a way to classify how efficient is an algorithm by means the execution time to solve a problem with the worst-case input), complexity of instances (it is computational complexity measures to determine the complex of the problems instances), and others elements. This work is focused on presenting a review of the theory of instances computational complexity. [ABSTRACT FROM AUTHOR]