改进的菌群算法研究

为了提高菌群寻优算法（ Bacterial Foraging Optimization, BFO）的搜索能力和解决多峰值复杂适应度函数模型避免过早收敛的问题，文中对原始菌群算法进行改进，提出多峰值菌群算法。将寻优过程分成两个时期，前期和原始菌群算法相同，在菌群收敛的后期，加入峰值数目和区间的判断，将区间编号，保证区间内部单峰值；然后在区间内部迭代运行菌群搜索，独立寻优，在多峰值和较复杂模型的情况下进行研究和评估。实验表明，在收敛速度、收敛稳定性和寻找全局最优方面均优于原始菌群算法。

Research on Optimized Bacterial Foraging Algorithm

To improve the search capabilities of bacterial foraging optimization andsolve the problem that the multiple peak complex adaptive function avoids premature convergence, a new type of multiple peak bacterial foraging optimization algorithm （MBFO） is proposed in this paper by improving bacterial foraging optimization. Divide the process into two parts, in the beginning of the algorithm, the operating is set as the same as the bacterial foraging optimization. In the second period, add a judgement of the number and the location of peak. Make sure only one peak in each area. And in the area do the iteration of the bacterial foraging optimization, looking for the optimal independently ,studying and estimating at the case of multiple peak and complex model. The experiment results indicate that the new type of multiple peak bacterial foraging optimization shows better results than bacterial foraging optimization in convergence rate, stability and looking for the global optimal.