BinEHO: a new binary variant based on elephant herding optimization algorithm
- PDF / 967,931 Bytes
- 21 Pages / 595.276 x 790.866 pts Page_size
- 64 Downloads / 161 Views
(0123456789().,-volV)(0123456789(). ,- volV)
ORIGINAL ARTICLE
BinEHO: a new binary variant based on elephant herding optimization algorithm Huseyin Hakli1 Received: 31 December 2019 / Accepted: 6 April 2020 Ó Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract One of the new optimization techniques proposed in recent years is elephant herding optimization (EHO) algorithm. Despite its short history, EHO has been used to solve many engineering and real-world problems by attracting researcher attention with its advantages such as efficient global search ability, having fewer control parameters and ease of implementation. However, there is no remarkable binary variant of EHO algorithm in the literature. A new binary approach based on EHO algorithm is proposed in this study. The newer binary variant of EHO named as BinEHO is binarized with preserving the search ability of basic EHO. The main purpose of the study is to present a simple, efficient and robust binary variant which copes with different binary problems. Therefore, the proposed method is tested on three important binary optimization problems, 0–1 knapsack, uncapacitated facility location and wind turbine placement, in order to show its performance and accuracy. In addition, the BinEHO is compared with various binary variants on these problems. Experimental results and comparisons show that the BinEHO algorithm is a robust and efficient tool for binary optimization. Keywords Binary optimization Elephant herding optimization Knapsack problem Uncapacitated facility location
1 Introduction Optimization problems can be seen everywhere in real life, including different areas such as engineering, finance, medicine, physics, manufacturing, chemistry and many others. These problems are classified based on various characteristics such as number of variables, type of variables, constraints, number of optima and degree of nonlinearity [1]. Considering the type of variables, optimization problems can be evaluated as two subareas: continuous and discrete. When variables of problem are xi 2 R ði ¼ 1; . . .; nÞ, the problem is referred as continuous; if xi 2 Z, the problem is named as discrete. Binary optimization is the most important branch of discrete optimization. In the binary optimization problems, each decision variable can be either 0 or 1 [2]. Most of the optimization problems can be regarded as a binary & Huseyin Hakli [email protected] 1
Department of Computer Engineering, Necmettin Erbakan University, 42090 Konya, Turkey
optimization problem [3]. These problems have numerous implementations in practice, so they attract attention of practitioners and researchers from various disciplines [4]. The unit commitment problem [5, 6], uncapacitated facility location problem [7, 8], 0–1 knapsack problem [9, 10], scheduling problem [11], wind turbine placement problem [12, 13], feature selection [14] are some common instances of binary optimization problems. Optimization methods based on swarm intelligence and evolutionary are used to find an a
Data Loading...
