基于折射反向学习的人工蜂群算法

Artificial Bee Colony Algorithm Based on Refracted Opposition-Based Learning

摘要: 人工蜂群算法(ABC)作为一种简单而有效的计算技术，被广泛应用于解决工程问题，为克服算法的局部搜索能力弱，收敛速度慢的缺点，提出了基于折射反向学习的人工蜂群算法(CRABC)。在引领蜂阶段，根据当前最优解指导产生下一次迭代的候选解；在引领蜂阶段后，计算当前种群的折射反向解，根据适应度值对种群择优选择组成下次迭代的候选解，极大提高了算法的收敛速度。为验证所提CRABC算法的性能，采用了12个基准测试函数进行测试，将CRABC算法与ABC算法，仅加入差分变异的ABC算法(CABC)，仅加入折射反向学习的ABC算法(RABC)进行比较，来验证综合两种策略后的CRABC算法改进效果；同时将CRABC算法，RABC算法与GABC算法，ABC/best/2算法在5个基准测试函数上进行比较。实验结果表明：所提的CRABC算法可以提高ABC算法的开发和探索能力。

Abstract: Artificial bee colony algorithm (ABC) as a simple and effective computing technology, is widely used to solve engineering problems, in order to overcome the drawbacks of the algorithm’s weak local search ability, slow convergence speed, Artificial bee colony algorithm based on refracted opposition-based learning (CRABC) was proposed. In the employed bee phase, the candidate solution of the next iteration is generated according to the current optimal solution. After the employed bee phase, calculate the refracted opposite solution of the current population, according to the fitness value, excellent individuals are selected to form the candidate solution of the next iteration, which greatly improves the convergence speed of the algorithm. To verify the optimization performance of the proposed algorithm, 12 benchmark functions were utilized to investigate the algorithm, the CRABC algorithm was compared with ABC algorithm, ABC algorithm with only differential variation (CABC), and ABC algorithm (RABC) with only refracted opposition-based learning to verify the improvement effect of the CRABC algorithm after the two strategies. Meanwhile, CRABC algorithm, RABC algorithm, GABC algorithm and ABC/best/2 algorithm are compared on five benchmark functions. The experimental results show that the proposed CRABC algorithm can improve the ability to develop and explore of the ABC algorithm.