高级检索
    毕晓君, 王 珏, 李 博, 李吉成. 基于动态迁移的ε约束生物地理学优化算法[J]. 计算机研究与发展, 2014, 51(3): 580-589.
    引用本文: 毕晓君, 王 珏, 李 博, 李吉成. 基于动态迁移的ε约束生物地理学优化算法[J]. 计算机研究与发展, 2014, 51(3): 580-589.
    Bi Xiaojun, Wang Jue, Li Bo, Li Jicheng. An ε Constrained Biogeography-Based Optimization with Dynamic Migration[J]. Journal of Computer Research and Development, 2014, 51(3): 580-589.
    Citation: Bi Xiaojun, Wang Jue, Li Bo, Li Jicheng. An ε Constrained Biogeography-Based Optimization with Dynamic Migration[J]. Journal of Computer Research and Development, 2014, 51(3): 580-589.

    基于动态迁移的ε约束生物地理学优化算法

    An ε Constrained Biogeography-Based Optimization with Dynamic Migration

    • 摘要: 提出基于动态迁移的ε约束生物地理学优化算法(εBBO-dm).首先,利用ε约束方法来处理约束条件,并根据群体约束违反度的优劣程度对水平参数ε进行自适应调整,充分利用较优不可行个体的有效信息,有效提高对可行域的搜索效率.其次,采用新的ε约束排序机制确定迁入率和迁出率,较好地平衡可行个体与不可行个体之间的关系.再次,为了增强迁移机制的搜索能力,提出新的动态迁移策略.最后,采用分段logistic混沌映射改进物种变异机制,提高了算法的收敛精度.通过对13个标准测试函数的仿真实验表明,εBBO-dm较其他算法在收敛精度和收敛速度上具有明显优势,尤其适合于复杂单目标约束优化问题的求解.

       

      Abstract: A new ε constrained biogeography-based optimization with dynamic migration, εBBO-dm, is proposed to solve constrained optimization problems. In the proposed algorithm, the ε constrained method is utilized to handle the constraints. According to the constraint violation of the colony, the ε level is set to utilize the useful information for the better infeasible individual sufficiently and to improve the search efficiency for the feasible space. Simultaneously, based on the feature of ε constrained method, a new ordering rule based on ε constrained is used to obtain the immigration rate and the emigration rate, which can dynamically balance the relation between the feasible individuals and the infeasible individuals. Additionally, a new dynamic migration strategy is shown to enhance the search ability of the migration mechanism. Eventually, with the purpose of improving the precision of convergence, the piecewise logistic chaotic map is introduced to improve the variation mechanism. Numerical experiments on 13 well-known benchmark test functions show that εBBO-dm is competitive with other optimization algorithms on the accuracy and the speed of convergence, especially when being applied to solve complex single-objective COPs.

       

    /

    返回文章
    返回