College of Mechanical and Electrical Engineering, Northeast Forestry University, Harbin 150040, China.
Math Biosci Eng. 2022 Aug 1;19(11):10963-11017. doi: 10.3934/mbe.2022512.
Aquila Optimizer (AO) and African Vultures Optimization Algorithm (AVOA) are two newly developed meta-heuristic algorithms that simulate several intelligent hunting behaviors of Aquila and African vulture in nature, respectively. AO has powerful global exploration capability, whereas its local exploitation phase is not stable enough. On the other hand, AVOA possesses promising exploitation capability but insufficient exploration mechanisms. Based on the characteristics of both algorithms, in this paper, we propose an improved hybrid AO and AVOA optimizer called IHAOAVOA to overcome the deficiencies in the single algorithm and provide higher-quality solutions for solving global optimization problems. First, the exploration phase of AO and the exploitation phase of AVOA are combined to retain the valuable search competence of each. Then, a new composite opposition-based learning (COBL) is designed to increase the population diversity and help the hybrid algorithm escape from the local optima. In addition, to more effectively guide the search process and balance the exploration and exploitation, the fitness-distance balance (FDB) selection strategy is introduced to modify the core position update formula. The performance of the proposed IHAOAVOA is comprehensively investigated and analyzed by comparing against the basic AO, AVOA, and six state-of-the-art algorithms on 23 classical benchmark functions and the IEEE CEC2019 test suite. Experimental results demonstrate that IHAOAVOA achieves superior solution accuracy, convergence speed, and local optima avoidance than other comparison methods on most test functions. Furthermore, the practicality of IHAOAVOA is highlighted by solving five engineering design problems. Our findings reveal that the proposed technique is also highly competitive and promising when addressing real-world optimization tasks. The source code of the IHAOAVOA is publicly available at https://doi.org/10.24433/CO.2373662.v1.
Aquila 优化器（AO）和非洲秃鹫优化算法（AVOA）是两种新开发的元启发式算法，分别模拟了自然界中鹰和非洲秃鹫的几种智能狩猎行为。AO 具有强大的全局探索能力，但局部开发阶段不够稳定。另一方面，AVOA 具有很好的开发能力，但探索机制不足。基于这两种算法的特点，本文提出了一种改进的混合 AO 和 AVOA 优化器，称为 IHAOAVOA，以克服单一算法的缺陷，为解决全局优化问题提供更高质量的解决方案。首先，将 AO 的探索阶段和 AVOA 的开发阶段结合起来，保留各自的有价值的搜索能力。然后，设计了一种新的复合基于对立的学习（COBL），以增加种群多样性，并帮助混合算法摆脱局部最优。此外，为了更有效地指导搜索过程，平衡探索和开发，引入了适应度距离平衡（FDB）选择策略来修改核心位置更新公式。通过在 23 个经典基准函数和 IEEE CEC2019 测试套件上与基本 AO、AVOA 和六个最先进的算法进行比较，全面研究和分析了所提出的 IHAOAVOA 的性能。实验结果表明，在所测试的大多数函数中，IHAOAVOA 比其他比较方法具有更好的求解精度、收敛速度和避免局部最优的能力。此外，通过解决五个工程设计问题，突出了 IHAOAVOA 的实用性。研究结果表明，该技术在解决实际优化任务时也具有很强的竞争力和潜力。IHAOAVOA 的源代码可在 https://doi.org/10.24433/CO.2373662.v1 上获得。
