求符号几何规划全局解的新方法  

A New Approach for Solving Golbal Solution of Signomial Geometric Programming

作　　者：申培萍[1] 李晓爱[1] 

机构地区：[1]河南师范大学数学与信息科学学院,新乡453007

出　　处：《工程数学学报》2006年第5期876-880,共5页Chinese Journal of Engineering Mathematics

基　　金：河南省自然科学基金(0511011500);河南省软科学研究计划(0513030920);河南省教育厅自然科学基金(2004110007)

摘　　要：符号几何规划(SGP)问题经常出现在工程设计和管理中。本文利用目标函数和约束函数的线性下界估计,提出一种求(SGP)问题全局解的线性松弛方法。与凸松弛方法相比,本文方法在计算上更加简单和容易,且在产生线性松弛的过程中没有引入新的变量和约束。数值例子表明所给方法是可行和有效的。Signomial geometric programming （SGP） problems often occur in engineering design and management. By utilizing the linear underestimation of objective and constraint functions, a linear relaxation method is proposed for finding global solution of SGP. In comparison with the convex relaxation method, this approach is more simple in computation, and it does not introduce new variables and constraints in forming linear programming relaxations, Numerical examples show the feasibility and efficiency of the proposed method.

关 键 词：符号几何规划 全局解 线性规划松弛 

分 类 号：O221.2[理学—运筹学与控制论]