基于混合生物地理学优化的多目标优化算法

提出一种基于混合生物地理学优化算法的多目标进化算法(multi-objective optimization based on hybrid biogeography based optimization， MOBBO)。针对生物地理学优化算法(biogeography based optimization， BBO)自身的机制，建立适用于BBO的多目标进化模型。在模型中，结合栖息地个体间的Pareto支配关系对栖息地适应度指数进行了重新定义；为了保持栖息地种群的分布性，提出一种新的基于动态距离矩阵的分布性保持机制；同时，根据多目标优化的特点，提出了新的自适应迁入迁出率确定方式，动态迁移策略及分段logistic混沌变异策略。通过对测试函数ZDT和DTLZ的仿真实验表明，与现有多种多目标优化算法相比，MOBBO在解集的收敛性和分布的均匀性上均有明显改善，能够有效且高效地进行复杂多目标优化问题的求解。

Multi-objective optimization based on hybrid biogeography based optimization

A new multi objective optimization based on hybrid biogeography based optimization (MOBBO) algorithm is proposed to solve multi-objective optimization problems. According to the evolutionary mechanism of BBO, the model of multi-objective evolutionary algorithm (MOEAs) which applies to BBO is built. In the model, the habitat suitability index, which combines with the Pareto dominance relation between the habitat individuals, is redefined. Moreover, a new mechanism based on the matrix of dynamic distance is set to maintain the distribution of population individuals. Simultaneously, according to the feature on multi-objective optimization, the self-adaptive method of determining the immigration rate and the emigration rate, dynamic migration strategy and the mutation strategy of piecewise logistic chaos are improved to achieve better convergence performance. Numerical experiments on ZDT and DTLZ test functions show that MOBBO is competitive with current other MOEAs on the convergence and the distribution, and is capable of solving the complex high dimensional multi-objective opti- mization problems (MOPs) more effectively and efficiently.