An advanced hybrid algorithm for nonlinear function optimization with real world applications

Abstract

This article presents an advanced hybrid algorithm (haDEPSO) for nonlinear constrained function optimization problem. It consists of the advised advanced differential evolution, that is, aDE (wherein a novel mutation, crossover, and selection strategy is introduced, to avoid premature convergence) and particle swarm optimization, namely, aPSO (in which novel gradually varying parameters is utilized, to escape stagnation).The proposed haDEPSO achieve better results as aDE and aPSO provides diverse convergence characteristic to the solution space. Moreover, in haDEPSO distinct population is merged with other in a predefined way, to balance between exploration and exploitation. Efficiency of the proposed hybrid haDEPSO along with its integrating component aDE and aPSO are inspected on IEEE CEC'2006 constrained benchmark functions and IEEE CEC'2011 real world problems. According to the comparison results, the proposed methods achieved better results than traditional DE and PSO with their hybrids as well as over many state-of-the-art algorithms.

Harris hawk optimization trained artificial neural network for anomaly based intrusion detection system

“…19 Many optimization algorithms have been developed and employed to carry out many optimization challenges. [20][21][22] The primary argument is that by utilizing appropriate mathematical modeling and techniques, we can overcome most real-life problems.…”

“…They have been deployed in recent decades for intrusion detection and show promise for replacing traditional training algorithms like BP 19 . Many optimization algorithms have been developed and employed to carry out many optimization challenges 20‐22 . The primary argument is that by utilizing appropriate mathematical modeling and techniques, we can overcome most real‐life problems.…”