《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (S1): 163-168.DOI: 10.11772/j.issn.1001-9081.2022081155

• 先进计算 • 上一篇    

基于蚁群数量动态调整的改进蚁群优化算法

白玮1(), 王成2, 王彩玲1, 詹熙1, 张磊1   

  1. 1.中国人民解放军陆军工程大学 指挥控制工程学院,南京 210014
    2.中国人民解放军 92403部队,福州 350007
  • 收稿日期:2022-08-07 修回日期:2022-10-22 接受日期:2022-11-03 发布日期:2023-07-04 出版日期:2023-06-30
  • 通讯作者: 白玮
  • 作者简介:白玮(1983—),男,河北张家口人,讲师,博士,主要研究方向:人工智能、网络智能管理.baiwei_lgdx@126.com
    王成(1983—),男,福建莆田人,硕士,主要研究方向:信息系统安全与管理
    王彩玲(1981—),女,山东临沂人,硕士,主要研究方向:人工智能算法优化
    詹熙(1999—),男,江西上饶人,硕士研究生,主要研究方向:数据工程、时间序列分析
    张磊(1989—),男,江西宜春人,博士研究生,主要研究方向:强化学习、数据工程。
  • 基金资助:
    陆军工程大学基础前沿科技创新项目(KYZYJQZL2008)

Improved ant colony optimization algorithm based on ant number dynamic adjustment

Wei BAI1(), Cheng WANG2, Cailing WANG1, Xi ZHAN1, Lei ZHANG1   

  1. 1.College of Command and Control Engineering,Army Engineering University of PLA,Nanjing Jiangsu 210014,China
    2.Unit 92403,PLA,Fuzhou Fujian 350007,China
  • Received:2022-08-07 Revised:2022-10-22 Accepted:2022-11-03 Online:2023-07-04 Published:2023-06-30
  • Contact: Wei BAI

摘要:

蚁群优化(ACO)算法是一种常用的元启发式算法,它通过模拟蚁群寻找食物的过程,为求解多维背包问题(MKP)等NP难(Non-deterministic Polynomial hard)问题提供可行途径。原始ACO算法及其改进算法,通常分为多个轮次,每个轮次均会生成一个蚂蚁种群寻找可行解。在不同轮次中,每轮蚁群中蚂蚁的数量是固定的,因此,如果将其指定一个较大的值,会导致算法出现不必要的时间消耗;反之,如果指定的值较小,则会降低算法全局最优解搜索能力。为此,提出了一种基于蚁群数量动态调整的改进蚁群优化算法ACO-ANDA(ACO algorithm based on Ant Number Dynamic Adjustment),所提算法在可行解搜索过程中,引入了一种新的蚁群数量动态调整机制。在每轮可行解搜索结束后,均根据近几轮可行解和历史最优解之间的关系,调整下一轮蚁群数量,实现对算法时间耗费和最优解搜索能力的平衡。再基于MKP基准测试集SAC-94的多组实验结果表明,相较于原始ACO算法,所提算法能够在最优解利润平均降低0.02%的情况下,平均降低77.85%的时间耗费。

关键词: 元启发式算法, 蚁群优化算法, 多维背包问题, 蚁群数量, 动态调整

Abstract:

The Ant Colony Optimization (ACO) algorithm is a common meta-heuristic algorithm, which provides feasible approaches to solve NP-hard (Non-deterministic Polynomial hard) problems such as Multi-dimensional Knapsack Problem (MKP) by simulating the process of ant colony searching for food. The original ACO algorithm and its improved algorithms are often divided into several rounds. In each round, a fixed number of ants are generated to find a feasible solution, which will lead to the difficulty of determining the number of ants, as a larger number will lead to unnecessary time consumption, and a smaller number will lead to a reduction in the optimal solution search capability. To tackle this issue, an improved ant colony optimization algorithm, named ACO-ANDA (ACO algorithm based on Ant Number Dynamic Adjustment) was proposed, which introduced a novel ant number dynamic adjustment mechanism. At the end of each round of feasible solution search, the ant number was adjusted for the next round based on the relationship between feasible solutions and historical optimal solutions in recent rounds to achieve a balance between time consumption and optimal solution search capability. Some experiments on the benchmark testing set of MKP SAC-94 were performed to compare with the original ACO algorithm. The results showed that the proposed algorithm could reduce the time consumption by 77.85% while the profit of the optimal solution was only reduced by 0.02% on average.

Key words: meta-heuristic algorithm, Ant Colony Optimization (ACO) algorithm, Multi-dimensional Knapsack Problem (MKP), ant number, dynamic adjustment

中图分类号: