Kusla Vipan, Brar Gurbinder Singh, Kaur Harpreet, Sandhu Ramandeep, Prabha Chander, Hassan Md Mehedi, Abdulla Shahab, Alam Md Rittique, Alshathri Samah, El-Shafai Walid
Department of Computer Science and Engineering, Sant Longowal Institute of Engineering and Technology, Longowal, Sangrur, Punjab, 148106, India.
School of Computer Science & Engineering, Lovely Professional University, Jalandhar, Punjab, 144411, India.
Sci Rep. 2025 Apr 22;15(1):13971. doi: 10.1038/s41598-025-97015-1.
Metaheuristic algorithms play a vital role in addressing a wide range of real-world problems by overcoming hardware and computational constraints. The Chameleon Swarm Algorithm (CSA) is a modern metaheuristic algorithm that uses how chameleons act. To improve the capabilities of the CSA, this work proposes a modified version of the Chameleon Swarm Algorithm to find better optimal solutions applicable to various application areas. The effectiveness of the proposed algorithm is assessed using 97 typical benchmark functions and three real-world engineering design problems. To validate the efficacy of the proposed algorithm, it has been compared to a number of well-known and widely-used classical algorithms, the Gravitational Search Algorithm, the Earthworm Optimization. The proposed modified Chameleon Swarm Algorithm using Morlet wavelet mutation and Lévy flight (mCSAMWL) is superior to existing algorithms for both unimodal and multimodal functions, as demonstrated by Friedman's mean rank test as well as three real world engineering design problems. Five performance metrics-average energy consumption, total energy consumption, total residual energy, dead node and cluster head frequency are taken into consideration when evaluating the performances against state-of-the-art algorithms. For nine different simulation scenarios, the proposed algorithm mCSAMWL outperforms the Atom Search Optimization (ASO), Hybrid Particle Swarm Optimization and Grey Wolf Optimization (PSO-GWO), Bald Eagle Search Algorithm (BES), the African Vulture Optimization Algorithm (AVOA), and the Chameleon Swarm Algorithm (CSA) in terms of average energy consumption and total energy consumption by 50.9%, 52.6%, 45%, 42.4%, 50.1% and 51.4%, 53.3%, 45.6%, 42.4%, 50.7%.
元启发式算法通过克服硬件和计算限制，在解决广泛的现实世界问题中发挥着至关重要的作用。变色龙群算法（CSA）是一种现代元启发式算法，它借鉴了变色龙的行为方式。为了提高CSA的性能，本文提出了一种改进的变色龙群算法，以找到适用于各种应用领域的更好的最优解。使用97个典型基准函数和三个实际工程设计问题来评估所提算法的有效性。为了验证所提算法的有效性，将其与一些著名且广泛使用的经典算法进行了比较，如引力搜索算法、蚯蚓优化算法。如Friedman平均秩检验以及三个实际工程设计问题所示，所提出的使用Morlet小波变异和Lévy飞行的改进变色龙群算法（mCSAMWL）在单峰和多峰函数方面均优于现有算法。在与最先进算法进行性能评估时，考虑了五个性能指标——平均能耗、总能耗、总剩余能量、死节点和簇头频率。对于九种不同的模拟场景，所提出的算法mCSAMWL在平均能耗和总能耗方面分别比原子搜索优化算法（ASO）、混合粒子群优化算法和灰狼优化算法（PSO - GWO）、秃鹰搜索算法（BES）、非洲秃鹫优化算法（AVOA）以及变色龙群算法（CSA）高出50.9%、52.6%、45%、42.4%、50.1%和51.4%、53.3%、45.6%、42.4%、50.7%。
