广义几何规划的加速全局优化算法  

An accelerating algorithm for solving global solution of generalized geometric programming

作　　者：王开荣[1] 马琳[1] 

机构地区：[1]重庆大学数学与统计学院,重庆401331

出　　处：《山东大学学报（理学版）》2013年第1期72-77,共6页Journal of Shandong University(Natural Science)

基　　金：重庆市高等教育教学改革研究重点项目(102104)

摘　　要：广义几何规划(generalized geometric programming,GGP)问题广泛出现在工程设计、风险管理以及工业制造等实际应用中。基于凸松弛提出GGP的加速全局优化算法,该算法通过新的剪枝技术,能将当前计算的不存在全局最优解的区域去除,从而加速算法收敛到全局最优解。最后给出了算法的收敛性证明。数值实验表明该算法是可行有效的。Generalized Geometric Programming（GGP） problems widely occur in engineering design, risk management, manufacturing, and so on. Based on the convex relaxation, an accelerating algorithm for solving global solution of gen- eralized geometric programming was proposed. A new pruning technique was used to cut away the current investigated region in which the global optimal solution does not exist and improve the convergence speed of this algorithm. Conver- gence of the algorithm was proved. Some experiments are reported to show the feasibility and efficiency of the proposed algorithm.

关 键 词：广义几何规划 全局优化 凸松弛 剪枝技术 

分 类 号：O221.2[理学—运筹学与控制论]