SONG Fei,WU Ze-zhong.A Comparative Study of External Penalty Function Method and Generalized Lagrangian Multiplier Method[J].Journal of Chengdu University of Information Technology,2017,(06):667-674.[doi:10.16836/j.cnki.jcuit.2017.06.016]
外罚函数法与广义Lagrange乘子法的比较研究
- Title:
- A Comparative Study of External Penalty Function Method and Generalized Lagrangian Multiplier Method
- 文章编号:
- 2096-1618(2017)06-0667-08
- 关键词:
- 约束优化; 外罚函数法; 广义Lagrange乘子法; 罚因子; 修正系数
- Keywords:
- constraint optimization; external penalty function method; generalized multiplier method; penalty factor; correction coefficient
- 分类号:
- O221.2
- 文献标志码:
- A
- 摘要:
- 基于非线性约束优化问题,讨论了外罚函数法与广义Lagrange乘子法,并通过MATLAB编程实现了两种算法。实验表明:(1)广义Lagrange乘子法在迭代次数和收敛结果上优于外罚函数法且对初始点的选取要求不高;(2)广义Lagrange乘子法的罚因子的修正系数不宜过大,一般在区间(1,2)上取值, 广义Lagrange乘子法更具优越性。最后,通过3个工业工程中的非线性规划实际问题说明乘子法比外罚函数法具有更广泛的实用性。
- Abstract:
- Based on the nonlinear constrained optimization problem,this paper discussed the external penalty function method and the generalized lagrangian multiplier method,and two algorithms have been implemented by programming.The experimental results show that:(1)The generalized multiplier method is superior to the external penalty function method in the iteration times and the convergence results,and the selection of the initial point is not strict.(2)The correction factor of the penalty factor of the generalized multiplier method should not be too large,evaluating on the interval(1,2)is more superior.Finally,the practical problems of nonlinear programming in three industrial projects show that the generalized lagrangian multiplier method has more practicability than the external penalty function method.
参考文献/References:
[1] Fiacco A V,McCormick G P.Nonlinear Programming:Sequential Unconstrained Minimization Techniques[J].Philadelphia:SIAM,1990.
[2] Courant R.Variational methods for the solution of problems with equilibrium and vibration[J].Bull.Amer.Math.Sco.,1943,49:1-23.
[3] Bertsekas D P.Constrained Optimization and Lagrange Multiplier Methods[M].New York:Academic Press,1982.
[4] Liebman J,Lasdon L S,Schrage L E,et al.Modeling and Optimaization with GINO[M].Redwood City:Science Press,1986.
[5] 刘兴高,胡云卿,李国栋,等.最优化方法应用分析[M].北京:科学出版社,2014.
[6] 倪勤.最优化方法与程序设计[M].北京:科学出版社.2009.
[7] 周密,尚利宏,胡瑜.给定余度下可重构计算系统的可靠性最优化研究[J].计算机科学,2009,36(4):83-298.
[8] 曾云宝,景旭,孙晓波.罚函数乘子的初值选取和多步长问题[J].哈尔滨理工大学学报,2003,8(1):20-22.
[9] 许苗.认知无线电系统频谱检测技术及其性能研究[D].上海:上海交通大学,2011.
备注/Memo
收稿日期:2017-06-27 基金项目:国家自然科学基金资助项目(71672013); 四川省软件科学研究计划资助项目( 2014ZR0016); 四川省哲学社会科学重点研究基地-系统科学与企业发展研究中心(重点)资助项目(Xq141306)