摘要
多流股换热器网络综合问题是一个混合整数非线性规划问题(M INLP),这类问题规模大、约束条件多,严重的非凸非线性使得目标函数存在多个局部最优解.传统的基于梯度的优化算法在求解时极易陷于局部最优.有鉴于此,本研究采用遗传算法解决此类问题,通过对遗传算法进行改进,针对简单遗传算法存在的早熟和运行参数难以确定的问题,设计了多样性保持算子和多种群进化的算法结构;计算时运行参数自适应确定,并把模拟退火算法思想引入遗传算法子代的生成中去.实例证明,采用所构造的算法可有效求解M INLP问题,并有利于寻求到全局最优解.
The synthesis of multi-stream heat exchanger network （MSHEN） is modeled as a mixed integer nonlinear programming problem （MINLP） that possesses large scale, many constraints and severe non-convexity and nonlinearity, thus resulting in many local optima. The traditional gradient-based optimization algorithms fail to find the global optimum. In order to solve this problem, the genetic algorithm （GA） is adopted in this paper. Moreover, to overcome the premature convergence and the difficulty in the appropriate determination of running parameters, an improved GA with diversity-maintaining operators and a multi-group evolution structure is proposed, in which the running parameters can be self-tuned and the simulated annealing algorithm is introduced during the evolution. Examples indicate that the adoption of the proposed algorithm helps solve the MINLP problem and guarantees the probability of finding the global optima in a MINLP problem.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2006年第8期6-12,共7页
Journal of South China University of Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(20536020)
关键词
多流股换热器网络
混合整数非线性规划
改进遗传算法
多样性保持算子
并行算法结构
multi-stream heat exchanger network
mixed integer nonlinear programming
improved genetic algorithm
diversity-retaining operator
parallel algorithm structure
