Artificial Bee Colony Algorithm Based on Adaptive Local Information Sharing Meets Multiple Dynamic Environments
Artificial Bee Colony Algorithm Based on Adaptive Local Information Sharing Meets Multiple Dynamic Environments
复制标题
DOI：
10.9746/jcmsi.12.1
复制
发表时间：
2019-01
期刊：
影响因子：
--
通讯作者：
R. Takano;Hiroyuki Sato;K. Takadama
中科院分区：
文献类型：
--
作者：
R. Takano;Hiroyuki Sato;K. Takadama
针对群优化方法中的阿尔蒂蜂群算法（ABC），通过对ABC算法的改进，提出了基于自适应局部信息共享的ABC算法（ABC-alis）。将ABC-alis应用于DOP中嵌入的各种类型的动态变化，以验证其在这种动态环境中的跟踪能力。具体地说，我们采用了以下五种动态变化和一种高维问题作为不同的环境：（A）局部最优值的周期性变化，（B）局部最优值的随机变化，（C）局部最优值坐标的随机变化，（D）两种随机变化的组合（B + C），（E）局部最优值的周期性变化，（E）局部最优值的周期性（E）环境（D）中局部最优的随机速度变化;以及（F）环境（E）中的高维问题。在这些实验中，以下三种方法进行了比较：ABC-alis作为建议的方法，ABC-lis作为我们以前的ABC-alis方法，基于物种的粒子群优化（SPSO）作为传统的方法。实验结果表明：（1）ABC-alis和ABC-lis在各种动态变化下都能比SPSO更快地捕捉到最优解，并保持更好的解;（2）在C、D、E环境下，ABC-alis从评价值、坐标、速度等方面都能适应局部最优解的随机变化;（3）ABC-alis即使在高维环境F中也能保持其性能。
: This paper focuses on the artificial bee colony (ABC) algorithm as one of swarm optimization methods and proposes ABC-alis (ABC algorithm based on adaptive local information sharing) by improving the ABC algorithm for dynamic optimization problems (DOPs). ABC-alis is applied to various types of dynamic changes embedded in DOPs to verify its tracking ability in such dynamic environments. Concretely, the following five types of dynamic changes and one of the high-dimensional problem are employed as the di ff erent environments: (A) a periodic change of evaluation values of local optima; (B) a random change of evaluation values of local optima; (C) a random change of local optima coordinates; (D) a combination of two kinds of random changes (B + C); (E) a random speed change of local optima in the environment (D); and (F) a high-dimensional problem in the environment (E). In these experiments, the following three methods are compared: ABC-alis as the proposed method, ABC-lis as our previous method of ABC-alis, and speciation-based particle swarm optimization (SPSO) as the conventional method. The experimental result revealed that the following implications: (1) ABC-alis and ABC-lis can capture the optimal solution more quickly and keep a better solution than SPSO in the various types of dynamic changes; (2) from environments C, D, and E, ABC-alis can adapt to the random change of local optima from the viewpoint of the evaluation value, coordinates, speed, and all of them; and (3) ABC-alis can maintain its performance even in the high dimensional environment F.