摘要
丙烯腈收率是丙烯腈装置的关键指标,如何得到丙烯腈收率是厂家很关注的研究,将新型优化算法用于丙烯腈收率软测量建模是1种较好的尝试。将新型微粒群优化算法用于同样新型的文化算法种群空间的优化,设计文化微粒群优化算法。它由种群空间和信念空间2部分组成,在种群空间和信念空间分别采用各自算法并行演化,同时,2个空间又根据一定的协议相互联系。分别将该算法和基本微粒群算法用于一些常用测试函数的优化问题;结果表明,与基本微粒群算法相比,文化微粒群算法加强了全局搜索能力,更容易收敛于全局最优解。最后将文化微粒群优化算法用于优化神经网络,构成文化微粒群神经网络,并将其应用于丙烯腈收率软测量建模。结果表明,此模型精度高,应用前景广阔。
During the process of acrylonitrile processing, acrylonitrile yield is a very important product indicator. Precisely obtaining the acrylonitrile yield is the key to control the process of acrylonitrile processing. It is a chanllege that some new optimization technologies are applied to construct the model of acrylonitrile yield. By combining particle swarm optimization algorithm(PSO) with cultural algorithm( CA), a new cultural particle swarm optimization algorithm(CPSO) is proposed in this paper. CPSO is constituted with the population space and the belief space and each space has its own optimization algorithm. Meanwhile, the two spaces communicate with each other by a communication agreement. Both CPSO and PSO were used to resolve the optimization problems of five widely used test functions, and the results show that CPSO has better performance than PSO. Then CPSO was applied in soft-sensor model of acrylonitrile yield. Experiment results show that the model based on CPSO has higher precision and better performance than the model based on PSO.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2010年第2期187-192,共6页
Computers and Applied Chemistry
基金
上海市教委重点学科(J51901)
上海市教委自然科学科研项目(06VZ002)资助
关键词
文化算法
微粒群算法
丙烯腈
优化
模型
cultural algorithm, particle swarm optimization algorithm, acrylonitrile , optimization, model