A particle swarm-BFGS algorithm for nonlinear programming problems

详细信息    查看全文

• 作者：Ali Mohammad Nezhad^a ; ^{a_mohammadnezhad@ie.sharif.edu} ; ^{ali.mohammadnezhad@gmail.com} ; Roohollah Aliakbari Shandiz^b ; ^{aliakbari_r@yahoo.com} ; Abdolhamid Eshraghniaye Jahromi^a ; ^{eshragh@sharif.edu}
• 关键词：Nonlinear programming ; Augmented Lagrangian penalty function ; BFGS algorithm ; Particle swarm optimization ; Active set SQP algorithm ; Interior/CG algorithm
• 刊名：Computers and Operations Research
• 出版年：2013
• 出版时间：April, 2013
• 年：2013
• 卷：40
• 期：4
• 页码：963-972
• 全文大小：340 K

文摘

This article proposes a hybrid optimization algorithm based on a modified BFGS and particle swarm optimization to solve medium scale nonlinear programs. The hybrid algorithm integrates the modified BFGS into particle swarm optimization to solve augmented Lagrangian penalty function. In doing so, the algorithm launches into a global search over the solution space while keeping a detailed exploration into the neighborhoods. To shed light on the merit of the algorithm, we provide a test bed consisting of 30 test problems to compare our algorithm against two of its variations along with two state-of-the-art nonlinear optimization algorithms. The numerical experiments illustrate that the proposed algorithm makes an effective use of hybrid framework when dealing with nonlinear equality constraints although its convergence cannot be guaranteed.