基于改进粒子群算法的工程项目综合优化  

Synthesis Optimization for Construction Project Based on Modified Particle Swarm Optimization Algorithm

作　　者：王维博[1,2] 冯全源[2] 

机构地区：[1]西华大学电气信息学院,四川成都610039 [2]西南交通大学信息科学与技术学院,四川成都610031

出　　处：《西南交通大学学报》2011年第1期76-83,共8页Journal of Southwest Jiaotong University

基　　金：国家自然科学基金-中国工程物理研究院联合基金资助项目(10876029);国家863计划资助项目(2009AA01Z230)

摘　　要：为解决现有粒子群优化算法进化过程中"早熟"的问题,提出了一种改进的粒子群优化算法HSPSO.算法采用多子群分层策略,以提高收敛速度和优化精度.为求解工程项目的综合优化问题,建立了工期-成本-质量的数学优化模型和多目标优化模型.通过实例对标准粒子群优化算法(SPSO)和差分进化(DE)算法进行了比较,并采用HSPSO算法进行多目标优化.最后,用枚举法验证了模型的合理性和算法的有效性.与已有研究相比,HSPSO算法能在种群规模较小(20个粒子)的情况下,快速找到满意的解(平均迭代次数不超过20次).A modified PSO（particle swarm optimization） algorithm—hierarchical subpopulation PSO（HSPSO） was proposed to avoid the premature phenomenon of the PSO algorithm during evolution.By using the strategy of subpopulation hierarchy,the algorithm can improve the convergence speed and accuracy.For the synthesis optimization of a construction project,mathematical optimization models and a multi-objective optimization model of construction time,cost and quality were established.In a case study,the standard PSO（SPSO） and differential evolution（DE） algorithms were compared,and the HSPSO algorithm was utilized to its synthesis optimization.In addition,the exhaustive enumeration was used to verify the effectiveness of these models and the feasibility of the HSPSO algorithm.The result shows that the HSPSO algorithm can quickly obtain satisfied results with average iterative times of less than 20 under the condition of a swarm size of 20 particles.

关 键 词：粒子群优化算法 项目管理 多目标 

分 类 号：O29[理学—应用数学]