创新多群果蝇优化算法与应用实例

130 浏览量更新于2024-08-25 收藏 1.55MB PDF 举报

本文主要探讨了一种新颖的多群果蝇优化算法，并将其应用于实际问题解决中。该研究发表在Elsevier出版社的一本期刊上，主要用于非商业性质的内部研究和教育用途，例如作者机构的教学和同事间的交流。然而，作者被明确禁止进行如复制、分发、销售或在个人、机构或第三方网站上发布文章的行为，除非符合Elsevier的版权和稿件政策。多群果蝇优化算法（Multi-Swarm Fruit Fly Optimization Algorithm, MSFFO）是一种基于模拟自然果蝇觅食行为的搜索算法。它借鉴了果蝇寻找食物时的群体协作和个体探索特性，将问题分解成多个子问题，每个子问题由一个果蝇群体独立处理，同时这些群体之间通过竞争与合作相互影响，以提高搜索效率和解决方案的质量。MSFFO算法的特点包括： 1. 群体结构：算法采用多群架构，每个群包含一组果蝇，它们各自执行独立的搜索路径，增强了全局搜索的灵活性。 2. 局部搜索与全局搜索：果蝇个体在局部区域进行精细搜索，而群体之间的信息交流则促进全局最优解的发现，避免陷入局部最优。 3. 适应性：根据果蝇的适应度值动态调整群体大小和飞行速度，以应对复杂问题的动态变化。 4. 随机性和进化：模仿果蝇的随机行为，如随机选择新的食物源或飞行方向，同时通过遗传操作更新种群，确保算法的进化能力。论文中，作者 Xiaofang Yu、Xiangshan Dai、Jingyi Zhao 和 Qian He 介绍了他们如何设计并实现这种算法，并展示了其在多个优化问题上的应用，比如工程设计、机器学习参数调优和数据挖掘等领域。实验结果显示，MSFFO算法具有良好的收敛性能和全局优化能力，能够在有限的时间内找到近似最优解。关键词：优化算法、果蝇优化、多群搜索、全局优化、工程优化、机器学习。通过这篇研究，读者不仅可以了解MSFFO的具体实现细节，还能学习到如何将类似的生物启发式算法应用于实际问题，提升计算效率和优化效果。这为未来的智能优化技术发展提供了新的思路和方法。

展开

Author's personal copy

One current trend in swarm behavior research is focusing more on the Drosophila (fruit ﬂy) species [11,12]. Recently, a

kind of Drosophila inspired optimization algorithm has been developed, called fruit ﬂy optimization algorithm (FOA) [13,14],

which is a novel evolutionary computation and optimization technique. The FOA is a new approach for ﬁnding global opti-

mization based on the swarm behavior of fruit ﬂy. The main inspiration of FOA is that the fruit ﬂy itself is superior to other

species in sensing and perception, especially in osphresis and vision. The FOA technique has the advantages of being easy to

understand and to be written into program code which is not too long. More recently, FOA technique has been applied in

several applications, such as swarms of mini autonomous surface vehicles [13] , neural network parameters optimization

[14–17], PID controller parameters tuning [18], key control characteristics optimization [19] etc. In order to improve the

search efﬁciency and global search ability, several researchers have also presented improved FOA [19–22]. However, for

these FOAs, fruit ﬂy swarm will use vision to ﬂy towards so far best smell position, this implies that fruit ﬂies will be around

the so far best smell position at a fast speed. The diversity loss occurs when the global optimal is shifted away from a too

converged swarm. This kind of swarm behavior is quite similar to being trapped in local optimal or premature in multi-

model optimization problems.

This paper presents a variation on the original FOA technique, named multi-swarm fruit ﬂy optimization algorithm

(MFOA), employing multi-swarms behavior to signiﬁcantly improve the performance. In the MFOA approach, several sub-

swarms moving independently in the search space with the aim of simultaneously exploring global optimal at the same

time, and local behavior between sub-swarms are also considered. In addition, several other improvements for original

FOA technique is also considered, such as: shrunk exploring radius using osphresis, and a new distance function. Application

of this new MFOA approach on several benchmark functions and parameter identiﬁcation of synchronous generator shows a

marked improvement in performance over original FOA technique.

The rest of this paper is organized as follows. Review of FOA technique is summarized in Section 2. Section 3 describes the

motivation and implement of the MFOA approach in detail. In Section 4, the testing of the proposed MFOA approach through

benchmark problems and parameter identiﬁcation of synchronous generator is carried out and the simulation results are

compared. Finally, the conclusion is drawn in Section 5.

2. FOA technique

2.1. Swarm behavior of fruit ﬂy

The osphresis organs of fruit ﬂy can ﬁnd all kinds of scents ﬂoating in the air; it can even smell food source from 40 km

away. Then, after it gets close to the food location, it can also use its sensitive vision to ﬁnd food and the company’s ﬂock-

ing location, and ﬂy towards that direction too [13]. When a ﬂy decides to go for hunting, it will ﬂy randomly to ﬁnd the

location guided by a particular odor. While searching, a ﬂy also sends and receives information from its neighbors and

makes comparison about the so far best location and ﬁtness [13]. If a ﬂy has found its favorable spot, it will then identify

the ﬁtness by taste. If the location no longer exists or the taste is ‘bitter’, the ﬂy will go off searching again. The ﬂy will

stay around at the most proﬁtable area, sending, receiving and comparing information with its swarm at the same time

[13].

The main idea behind the FOA te chnique is based upon th e D rosophil a’s biological behavi or [13]:(1)Theﬂyﬂies

with Levy ﬂight motio n; (2) It smells the potential location (attractiveness); (3) It would then taste. If it is n ot to its

liking (ﬁtness/proﬁtability), it rejects a nd goes to anot her location. To the ﬂy, attract iveness is not necessarily prof-

itable; (4) While foraging or mating, the ﬂy also sends and receives messages with its swarm about its food and their

mates.

2.2. FOA technique implementation

Based on the food ﬁnding characteristics of fruit ﬂy swarm, a kind of FOA technique is proposed in [13,14], which is a

novel evolutionary computation and optimization technique. Although the FOAs are inspired by swarm behavior of fruit

ﬂy, the implement procedure of FOA in [13] is different from that in [14–18]. This can be considered as two different ways

of computing implement for FOA technique. In this paper, we will focus on FOA technique in [14–18] for improvement. The

FOA in [14–18] can be divided into several necessary steps and the main steps are described as follows:

Step 1. Random initial fruit ﬂy swarm location as shown in Fig. 1. Init X



axis; Init Y



axis.

Step 2. Give the random direction and distance for the search of food using osphresis by an individual fruit ﬂy.

¼ X



axis þ RandomValue

¼ Y



axis þ RandomValue ð1Þ

where i is the population size of fruit ﬂies.

Step 3. Since the food location cannot be known, the distance to the origin is thus estimated ﬁrst (Dist), then the smell

concentration judgment value (S) is calculated, and this value is the reciprocal of distance.

X. Yuan et al. / Applied Mathematics and Computation 233 (2014) 260–271

261

下载后可阅读完整内容，剩余12页未读，立即下载

身份认证购VIP最低享 7 折!

30元优惠券

weixin_38545485

粉丝: 5

创新多群果蝇优化算法与应用实例

双子群果蝇优化算法解决0-1背包问题

果蝇优化算法：一种创新的优化技术

基于混沌粒子群的果蝇优化算法研究

【智能优化算法】果蝇优化算法（FOA）附Python代码.zip

果蝇优化算法FOAGRNN_Training.txt

改进步长与策略的果蝇优化算法

基于优化算法的核函数参数选择的研究

果蝇优化算法新进展：Modified FOA与金融模型应用

MATLAB优化BP神经网络的果蝇算法实现及预测分类

改进果蝇算法Matlab实现及源码下载

最新资源