Читать книгу Nature-Inspired Algorithms and Applications - Группа авторов - Страница 39

Elephant Herding Optimization (EHO) algorithm is one of the metaheuristic approach swarm-based search algorithms that is utilized to explain various problems of optimization and also utilized benchmark, localization based on energy, services selection in QOS web service compositions, appliance scheduling in smart grid identification, and PID controller tuning– based problems. The algorithm is inspired by the performance of group of elephant in the wild, in which elephants live in a group with a female elephant called leader matriarch, while the male are disconnected from the group when they are adulthood. The EHO algorithm is based on the models of collecting behaviors of elephants in two procedures. They are clan update and separation. Clan update is referred as updating the elephants and matriarch present location in every clan and separation is referred as enhancing the populace range in the subsequent phase of search.

Table 1.1 List of applications of various algorithms.

S. no.	Algorithm	Areas of application
1.	Memetic algorithm	Multi-dimensional knapsack problem, pattern recognition, feature/gene selection, training of artificial neural networks, clustering of gene expression profiles, traveling salesman problem, Robotic motion planning
2.	Genetic algorithm	Allocation of document for a distributed system, PC robotized plan, server farm/server center, code breaking, criminological science, robot behavior, PC design, Bayesian inference, AI, game hypothesis
3.	Ant colony optimization algorithm	Problems of generalized assignment and the set covering, classification problems, Ant Net for organized directing and multiple knapsack Problem
4.	Particle swarm optimization algorithm	Combination with a back engendering calculation, to prepare a neural system framework structure, multi-target optimization, classification, image clustering and image clustering, image processing, automated applications, dynamic, pattern recognition, image segmentation, robotic applications, time frequency analysis, decision-making, simulation, and identification
5.	Harmony search algorithm	Power systems, power systems, transportation, medical science and robotics, industry and signal and image processing
6.	Artificial bee colony algorithm	Problem of medical pattern classification, network reconfiguration, minimum spanning tree, train neural networks, radial distribution system of network reconfiguration, and train neural networks
7.	Firefly algorithm	Semantic web composition, classification and clustering problems, neural network, fault detection, digital image compression, feature selection, digital image processing, scheduling problems, and traveling salesman problem
8.	Bat algorithm	Image processing, clustering, classification, data mining, continuous optimization, problem inverse and estimation of parameter, combination scheduling and optimization, and fuzzy logic

The working of EHO is based on that every elephant in clan is updated by utilizing group data through clan by the procedure of updating, and afterward, the poorest elephant is supplanted by randomly produced elephant individual through the procedure of updating. EHO can discover much improved solutions on more problems of benchmark. Problems of benchmark are a lot of different types of problem of optimization that comprises of different kinds of aptitudes that utilized in testing and the estimation is verified and described. Then, the execution of estimation enhances the algorithm under various ecological conditions.

Table 1.1 lists the various applications of NIC algorithms.