计算机工程 ›› 2011, Vol. 37 ›› Issue (8): 161-163.doi: 10.3969/j.issn.1000-3428.2011.08.055

• 人工智能及识别技术 • 上一篇    下一篇

交通网络设计问题的人工鱼群算法

刘炳全,孙广才   

  1. (渭南师范学院数学与信息科学系,陕西 渭南 714000)
  • 出版日期:2011-04-20 发布日期:2012-10-31
  • 作者简介:刘炳全(1980-),男,讲师、硕士,主研方向:系统优化建模;孙广才,教授、博士
  • 基金项目:
    渭南师范学院科研基金资助重点项目(09YKF003)

Artificial Fish Swarm Algorithm for Traffic Network Design Problem

LIU Bing-quan, SUN Guang-cai   

  1. (Department of Mathematics and Information Science, Weinan Teachers University, Weinan 714000, China)
  • Online:2011-04-20 Published:2012-10-31

摘要: 人工鱼群算法是一种新的群体智能优化算法,可较好地避免局部极值并取得全局极值,但针对离散优化问题却存在开发平衡及探索能力差等缺点。为此,设计一种自适应变异的人工鱼群算法,在迭代过程中添加变异算子并自动调节视野范围和拥挤度因子。将该算法应用于多等级选择的离散型交通网络二层规划模型设计中,上下层模型分别采用人工鱼群算法及Frank-Wolfe算法进行求解,从而为求解这类模型提供新方法。仿真结果表明,该算法具有较好的稳定性和收敛速度,能够应用于大型城市交通网络设计中。

关键词: 人工鱼群算, 变异算子, 交通网络设计, 二层规划, 收敛速度

Abstract: Artificial Fish Swarm Algorithm(AFSA) is a novel intelligence optimization algorithm, which may effectively avoid local optimum solution and find the global optimum solution. But the algorithm has several disadvantages such as the poor ability to keep the balance of exploration and exploration for the discrete optimization problems. To overcome the problems, an algorithm with normal mutation operator is proposed. It can dynamically adjust the vision and congestion operator. In the bi-level programming model for discrete network design problems, the upper model is solved by AFSA with normal mutation operator and lower model by the Frank-Wolfe algorithm. Simulation results demonstrate this algorithm is efficient and effective, and it is suitable for applying to large scale road network.

Key words: Artificial Fish Swarm Algorithm(AFSA), mutation operator, traffic network design, bi-level programming, convergence rate

中图分类号: