Scholarlink

<p>The proposed algorithm is based on the combination of enhanced seagull optimization (ESO) algorithm, differential evolution (DE) algorithm, wild horse optimization (WHO) algorithm with probability matrix to solve optimization problems. The seagull optimized algorithm is enhanced to avoid premature convergence. A probabilistic matrix with an equal value of each location is generated having 3 columns and the number of populations as rows. The three columns represent enhanced seagull optimization algorithm, differential evolution algorithm, and wild horse optimization algorithm respectively. Initial decision vector is generated in the bound. The probabilistic matrix after each iteration is updated with respect to minimum fitness value among 3 algorithms for each population. The decision vector for the next stage is decided based on the highest value of probability matrix. This stage improves the learning process and reduces the convergence rate. The updation of the probabilistic matrix and selection of decision vector for the next stage is continued until certain criteria are achieved. The optimized result of the proposed algorithm is the final minimum value. The proposed methodology is validated by comparing mean, standard deviation, best and worst value with ten other algorithms for 95 functions of CEC-2021. These algorithms are Archimedes optimization, wild horse optimizer, cuckoo search, differential evolution, black widow optimization, seagull optimization, whale optimization, arithmetic optimization, coot optimization, enterprise development (ED), hybrid jellyfish search and particle swarm optimization (HJPSO), modified lightning search algorithm (MLSA), polar lights optimizer (PLO), and parrot optimizer algorithm. The efficiency of the proposed methodology was also validated statistically with the ten methodologies The proposed methodology is also verified by five real-time engineering optimization problems. The results show better performance of the proposed algorithm compared to other algorithms.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel hybrid ESO-DE-WHO algorithm for solving real-engineering optimization problems

  • 期刊论文

  • 发表日期：

    2025年1月14日

查看全文
  • Damodar Panigrahy,
  • Padarbinda Samal

摘要

The proposed algorithm is based on the combination of enhanced seagull optimization (ESO) algorithm, differential evolution (DE) algorithm, wild horse optimization (WHO) algorithm with probability matrix to solve optimization problems. The seagull optimized algorithm is enhanced to avoid premature convergence. A probabilistic matrix with an equal value of each location is generated having 3 columns and the number of populations as rows. The three columns represent enhanced seagull optimization algorithm, differential evolution algorithm, and wild horse optimization algorithm respectively. Initial decision vector is generated in the bound. The probabilistic matrix after each iteration is updated with respect to minimum fitness value among 3 algorithms for each population. The decision vector for the next stage is decided based on the highest value of probability matrix. This stage improves the learning process and reduces the convergence rate. The updation of the probabilistic matrix and selection of decision vector for the next stage is continued until certain criteria are achieved. The optimized result of the proposed algorithm is the final minimum value. The proposed methodology is validated by comparing mean, standard deviation, best and worst value with ten other algorithms for 95 functions of CEC-2021. These algorithms are Archimedes optimization, wild horse optimizer, cuckoo search, differential evolution, black widow optimization, seagull optimization, whale optimization, arithmetic optimization, coot optimization, enterprise development (ED), hybrid jellyfish search and particle swarm optimization (HJPSO), modified lightning search algorithm (MLSA), polar lights optimizer (PLO), and parrot optimizer algorithm. The efficiency of the proposed methodology was also validated statistically with the ten methodologies The proposed methodology is also verified by five real-time engineering optimization problems. The results show better performance of the proposed algorithm compared to other algorithms.