提升效率的自适应统一人工蜂群算法优化策略
50 浏览量
更新于2024-08-28
收藏 245KB PDF 举报
全球优化的自适应统一人工蜂群算法（Adaptive Uniform Artificial Bee Colony, AUBBC）是一种基于生物启发式优化的计算方法，灵感来源于蜜蜂群体觅食行为。自2005年Karaboga提出Artificial Bee Colony (ABC)以来，由于其简单易实现且在众多领域如科学、工程、经济和社会生活中的广泛应用，它已经成为一种有效的优化工具，尤其是在与遗传算法（GA）、粒子 swarm optimization (PSO)、蚁群优化（ACO）和生物地理优化（BBO）等传统算法比较时，展现出了良好的性能。
然而，标准ABC算法也存在一些局限性。其中显著的问题是它的收敛速度相对较慢。与差分进化（DE）和粒子 swarm 这样的优化算法相比，ABC在寻找最优解的过程中，候选解生成策略可能不够高效，导致搜索过程中的迭代效率不高。这可能是由于ABC在模仿蜜蜂社会中，工蜂通过“舞蹈”传递信息的方式，可能存在局部搜索过度而忽视全局最优的可能性。
为解决这些问题，AUBBC算法旨在改进ABC的不足，通过引入自适应机制来增强算法的全局探索能力。具体来说，AUBBC可能采取以下策略：
1. **自适应参数调整**：根据当前搜索状态，动态调整种群大小、个体步长等关键参数，以更好地平衡局部搜索和全局搜索的频率。
2. **混合策略**：结合多种优化技术，例如局部搜索或邻域搜索，以提高在局部最优区域跳出的能力。
3. **多峰搜索优化**：针对复杂问题的多模态特性，设计多中心搜索机制，同时处理多个潜在最优区域。
4. **学习与记忆机制**：通过学习历史搜索经验，存储和利用先前的有效解决方案，以避免重复劳动并加快收敛速度。
5. **进化算子改进**：对候选解产生和更新规则进行创新，如使用更高效的编码方式或适应性选择操作，以提升算法的整体性能。
全球优化的自适应统一人工蜂群算法（AUBBC）通过增强算法的动态性和灵活性，旨在克服标准ABC的收敛速度问题，以提供更高效的全局优化解决方案，尤其适用于那些需要快速找到全局最优解或者具有复杂搜索空间的应用场景。
1 INTRODUCTION
Several kinds of biological-inspired optimization
algorithms, such as genetic algorithm (GA) inspired by
the Darwinian law of survival of the fittest [1,2], particle
swarm optimization (PSO) inspired by the social
behavior of bird flocking or fish schooling [3,4], ant
colony optimization (ACO) inspired by the foraging
behavior of ant colonies [5], and biogeography-based
optimization (BBO) inspired by the migration behavior
of island species [6], have sprung up over the past
decades. Artificial bee colony (ABC) is such a
computation method proposed in 2005 by Karaboga
who is motivated by the foraging behaviors of honey bee
swarms [7]. In a number of studies, the artificial bee
colony algorithm performed effectively in comparison
to other stochastic optimization methods. It has been
widely used in a variety of applications including
science, engineering, economics and social life thanks
to its simplicity and ease of implementation.
However, some shortcomings of the standard ABC
are also exposed. For example, compared to DE and
PSO, the convergence speed of ABC is especially
slower owing to its candidate producing strategy. In
addition, it’s easy for ABC to get stuck on local when
dealing with complicated multimodal problems. It has
long been a question for a population-based
optimization algorithm to strike a balance between
exploration and exploitation because these two factors
contradict each other. While the standard ABC has a
great ability to explore in the unknown areas of the
solution space, the exploitation ability to utilize the
previous searching results is poor which may lead to
missing of the most promising search direction. The
above two defects limit the wide utilization of ABC
algorithm.
Therefore, great efforts have been made for a better,
modified ABC algorithm. [8] invented interactive ABC
which introduced the concept of universal gravitation
into the consideration of the affection between
employed bees and the onlooker bees. By incorporating
the information of global best (gbest) solution into the
solution search equation to take advantage of the
exploitation, [9] proposed gbest-guided ABC in 2010.
[10] used chaotic maps simulating the chaos in real bee
colony behavior for parameter adaption in population
initialization in order to improve the convergence speed
and prevent the ABC to get stuck on local solutions. [11]
presented another gbest-guided search equation and get
a new search mechanism by means of introducing a
selective probability P and employ chaotic systems as
well as opposition-based learning methods. Another
improvement in the ABC algorithm is controlling the
frequency of perturbation and the ratio of the variance
operator, proposed by [12]. In spite of these
modifications, it still seems to be a thorny problem to
get satisfied results in both aspects.
In this work, an adaptive unified ABC algorithm,
auABC for short, is proposed. Its search solution
equation retains the advantage of the exploration ability
while the global best (gbest) solution is also applied in
order to improve the exploitation. It employs five self-
adaptive parameters, all of which changes with iteration
times. During the prophase of iteration, exploration
work is mainly employed. WWhile during the later
iterative period, exploitation work is the focal point. In
addition, the new modified ABC uses a chaotic strategy
for parameter adaption to take charge of the proportion
of exploration and exploitation.
The rest of this paper is organized as follows.
Section 2 gives a summary of ABC algorithm. The
proposed adaptive unified ABC algorithm is presented
and analyzed in section 3. In section 4, the experiments
are carried out and the results are discussed by
comparison. In section 5, conclusions are provided.
An Adaptive Unified Artificial Bee Colony Algorithm for Global Optimization
Yang Yang
1
, Feiyi Xu
1
, Haidong Hu
2
, Hao Gao
1*
1.
The Institute of Advanced Technology, Nanjing University of Posts and Telecommunications, Nanjing, China
2.
Beijing Institute of Control Engineering, Beijing, China
E-mail: tsgaohao@gmail.com
Abstract: The artificial bee colony (ABC) algorithm invented by Karaboga is a relative new nature inspired heuristic for
optimization problem. It has been proved to be competitive with some conventional optimization algorithms. This paper
proposes an adaptive unified artificial bee colony (auABC) algorithm which employs a single equation unifying multiple
strategies into one expression. As we all know, ABC is good at exploration but poor at exploitation due to the
insufficiency in its solution search equation and its one-dimensional search strategy leads to slower convergence speed.
In order to improve the above defects, we created a combined self-adaptive equation that its parameters are determined
by the current iteration times to solve the problem mentioned above. We also introduce a chaotic strategy when
generating candidate food sources which balance the proportion of exploration and exploitation. Experiments conducted
on benchmark functions demonstrate that our algorithm achieves good performance in both unimodal and multimodal
functions as expected compared to several state-of-the-art algorithms.
Key Words: Artificial bee colony algorithm, Solution search equation, Adaptive strategy
5497
978-1-5386-1243-9/18/$31.00
c
2018 IEEE
下载后可阅读完整内容，剩余5页未读，立即下载
2021-03-06 上传
2018-03-20 上传
2021-04-13 上传
点击了解资源详情
点击了解资源详情
点击了解资源详情
点击了解资源详情
点击了解资源详情
2022-07-15 上传
