摘要
在实际的化工过程中会遇到许多非线性优化问题。常规群智能优化算法在解决这类问题时,常出现收敛精度差和容易陷入局部最优,本文针对此提出了一种基于寄生行为的双种群萤火虫算法(FAPB)。该算法将进化种群均分为两个种群,通过生物的寄生行为将两个种群联系起来,共享进化信息,提高了全局搜索能力;为防止算法陷入局部最优,引入基于自适应系数的高斯变异机制,提高了局部搜索能力。对4个经典测试函数进行仿真,结果表明:与标准FA算法、FALS算法、LDPSO算法比较,FAPB算法在收敛精度和全局搜索能力上都有较大提升。将该算法应用于柴油调合过程,结果验证了其在实际应用中的可行性。
There exist many nonlinear optimization problems in the actual chemical process, for which the conventional swarm intelligence optimization algorithm easily falls into local o p t i m u m . This paper presents a parasitic behavior （FAPB ） based double population firefly algorithm. B y dividing into the evolutionary population into t w o ones linked together via the parasitic behavior of organisms, these population can share the information and improve the global searching ability. In order to prevent the algorithm from falling into local o p t i m u m , this paper further introduces Gauss mutation m e c h a n i s m based on the adaptive coefficient to improve the local search ability. B y simulations on four classical test functions, it is shown that, compared with the standard F A algorithm, F A L S algorithm and L D P S O algorithm, the proposed FAPB algorithm can effectively improve the convergence accuracy and global search capability. Finally,the feasibility of the proposed algorithm is also demonstrated via diesel blending.
出处
《华东理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2017年第2期213-219,共7页
Journal of East China University of Science and Technology
基金
国家自然科学基金重点项目(61333010)
国家自然科学基金面上项目(21376077
61403141)
上海市"科技创新行动计划"研发平台建设项目(13DZ2295300)
关键词
萤火虫算法
寄生行为
局部搜索
高斯变异
柴油调合
firefly algorithm
parasitic behavior
local search
Gauss variation
diesel blending
