线性分式规划的Frank-wolfe优化算法  

Frank-wolfe Optimization Algorithm of Linear Fractional Programming

作　　者：陈诚[1] 杨森[2] 李粉红[3] 

机构地区：[1]西安通信学院基础部,陕西西安710106 [2]河南科技大学理学院,河南洛阳471003 [3]商洛学院数学与计算科学系,陕西商洛726000

出　　处：《商洛学院学报》2010年第2期20-21,共2页Journal of Shangluo University

基　　金：商洛学院科研基金项目(08SKY031)

摘　　要：线性分式规划在经济管理领域有着广泛的应用,但分式规划是非线性优化模型,相对复杂。利用Frank-wolfe优化方法,将带有线性约束条件的一类线性分式规划转化为线性规划,求得此类线性分式规划的局部最优解;同时给出了算法的步骤,讨论了收敛性定理;通过数值实验,发现该算法具有迭代次数少等优点。The linear fractional programming was widely applied to economy and management field. But the linear fractional programming is nonlinear optmization model and relatively complex.Applying Frank-wolfe algorithm, the linear fractional programming with the linear constraint condition was transformed into linear programming. And the computational steps in determining all optimal solutions are given： convergence theorem is also discussed. Finally, numerial computations show that this algorithm has less iterative steps.

关 键 词：分式规划 线性化方法 收敛性 

分 类 号：O224[理学—运筹学与控制论]