A solution to bi/tri-level programming problems using particle swarm optimization

详细信息    查看全文

• 作者：Jialin Han^a ; ^b ; ^{Jialin.Han@uts.edu.au" class="auth_mail" title="E-mail the corresponding author} ; Guangquan Zhang^a ; ^{Guangquan.Zhang@uts.edu.au" class="auth_mail" title="E-mail the corresponding author} ; Yaoguang Hu^b ; ^{hyg@bit.edu.cn" class="auth_mail" title="E-mail the corresponding author} ; Jie Lu^a ; ^{Jie.Lu@uts.edu.au" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Bi-level programming ; Tri-level programming ; Multilevel decision-making ; Particle swarm optimization ; Computational intelligence
• 刊名：Information Sciences
• 出版年：2016
• 出版时间：20 November 2016
• 年：2016
• 卷：370-371
• 期：Complete
• 页码：519-537
• 全文大小：884 K

文摘

Multilevel (including bi-level and tri-level) programming aims to solve decentralized decision-making problems that feature interactive decision entities distributed throughout a hierarchical organization. Since the multilevel programming problem is strongly NP-hard and traditional exact algorithmic approaches lack efficiency, heuristics-based particle swarm optimization (PSO) algorithms have been used to generate an alternative for solving such problems. However, the existing PSO algorithms are limited to solving linear or small-scale bi-level programming problems. This paper first develops a novel bi-level PSO algorithm to solve general bi-level programs involving nonlinear and large-scale problems. It then proposes a tri-level PSO algorithm for handling tri-level programming problems that are more challenging than bi-level programs and have not been well solved by existing algorithms. For the sake of exploring the algorithms' performance, the proposed bi/tri-level PSO algorithms are applied to solve 62 benchmark problems and 810 large-scale problems which are randomly constructed. The computational results and comparison with other algorithms clearly illustrate the effectiveness of the proposed PSO algorithms in solving bi-level and tri-level programming problems.