计算机科学与探索 ›› 2023, Vol. 17 ›› Issue (12): 2913-2927.DOI: 10.3778/j.issn.1673-9418.2207004

• 理论·算法 • 上一篇    下一篇

角逐和信息素引导的多目标黑寡妇优化算法

傅彦铭,许励强,祁康恒,沈煜鸣,屈迟文   

  1. 1. 广西大学 计算机与电子信息学院,南宁 530004
    2. 右江民族医学院 公共卫生与管理学院,广西 百色 533000
  • 出版日期:2023-12-01 发布日期:2023-12-01

Multi-objective Black Widow Algorithm Guided by Competitive Mechanism and Pheromone Mechanism

FU Yanming, XU Liqiang, QI Kangheng, SHEN Yuming, QU Chiwen   

  1. 1. School of Computer and Electronic Information, Guangxi University, Nanning 530004, China
    2. School of Public Health and Management, Youjiang Medical College for Nationalities, Baise, Guangxi 533000, China
  • Online:2023-12-01 Published:2023-12-01

摘要: 黑寡妇优化算法(BWOA)是一种群体智能优化算法,具有收敛速度快、收敛精度高等优点。但BWOA所采用的更新策略过于简单,容易陷入局部最优解;其次在多维空间中搜索能力欠缺,种群结构单一,算法的收敛性和多样性有待改善。为提高BWOA的综合性能,并使其能够应用于多目标优化问题,提出一种角逐机制和改进信息素机制引导的多目标黑寡妇优化算法(MBWOA)。MBWOA采用动态分配种群的方法,在迭代过程中将种群一分为二,分别使用不同的角逐机制,增强迭代过程中种群的多样性,提升算法的收敛性;同时,使用改进的信息素机制对经过角逐的子代个体进行更新,引导个体向种群间隙方向优化,改善种群的分布,增强算法的收敛能力。MBWOA与四个对比算法在IGD、HV、Spread三个指标上分别进行对比实验,结果表明MBWOA具有更好的收敛精度、收敛速度和多样性。最后,通过在三个指标上对MBWOA所用机制的对比实验,证实了所用机制的有效性。

关键词: 多目标优化, 黑寡妇优化算法(BWOA), 角逐机制, 改进信息素机制

Abstract: Black widow optimization algorithm (BWOA) is a swarm intelligence optimization algorithm, which has the advantages of fast convergence and high precision. However, the update strategy adopted by BWOA is too simple, and it is easy to fall into the local optimal solution. Moreover, the search ability in multi-dimensional space is lacking, the population structure is single, and the convergence and diversity of the algorithm need to be improved.  In order to improve the comprehensive performance of BWOA and make it applicable to multi-objective optimization problems, this paper proposes a multi-objective black widow optimization algorithm (MBWOA) guided by a competition mechanism and an improved pheromone mechanism. MBWOA adopts the method of dynamic allocation of populations, which divides the populations into two in the iterative process and uses different competition mechanisms to enhance the diversity of the populations in the iterative process and improve the convergence of the algorithm. At the same time, it uses the improved pheromone mechanism to guide offspring individuals that have gone through the competition mechanism to optimize in the direction of population gap, improve the distribution of population, and enhance the convergence ability of the algorithm. Using MBWOA and four comparison algorithms to conduct comparative experiments on three indicators of IGD, HV and Spread respectively, the results show that MBWOA has better convergence accuracy, convergence speed and diversity. Finally, the effectiveness of the used mechanism is confirmed by the experiments of MBWOA and the comparison algorithms on three indicators.

Key words: multi-objective optimization, black widow optimization algorithm (BWOA), competition mechanism, improved pheromone mechanism