An efficient opposition based Levy Flight Antlion optimizer for optimization problems

摘要

This work proposes a new efficient version of recently proposed Antlion Optimizer (ALO) namely Opposition based Levy Flight Antlion optimizer (OB-LF-ALO). The upgraded version is conceptualized on the theory of Opposition based learning integrated with Levy Flight for random walk in place of uniform distributed random walk in original ALO. The success of any optimization algorithm relies on adequate balancing of exploration and exploitation during the process of evolution. The original algorithm is prone to stagnate in local optima and requires diversified exploration with appropriate blending of exploitation. The proposed technique is well capable of accelerating convergence by enhancing the initial diversification and good exploitation capability at later stage of generations. The performance of developed algorithm is validated by applying a wide-ranging set of 27 unconstrained continuous benchmark test functions. The impact of generated random numbers after employing levy flight and updated population after applying opposition based learning during evolution is analysed using certain metrics such as trajectories, elite convergence curve, average of absolute distance between search agents before and after improving the algorithm and data distribution using box plot diagrams. A non-parametric Wilcoxon ranksum test is used to exhibit its statistical significance. The projected algorithm is also compared with its recently developed version namely opposition based laplacian antlion optimizer (OB-L-ALO). The algorithm is also established with wide range of real life classical engineering optimization problem including two unconstrained and two constrained problems. The experimental analysis establishes that the developed variant OB-LF-ALO is superior as compared to ALO and OB-L-ALO. (C) 2018 Elsevier B.V. All rights reserved.