基于反凸规划的两层线性规划问题全局最优解算法  

A GLOBAL OPTIMAL SOLUTION ALGORITHM FOR TWO-LEVEL LINEAR PROGRAMS BASED ON THE REVERSE CONVEX PROGRAMMING

作　　者：许丙胜[1,2] 邵建峰[3] 

机构地区：[1]南京工业大学经济管理学院,南京210009 [2]南京工业大学理学院,南京210009 [3]南京工业大学管理学院,南京210009

出　　处：《南京大学学报（数学半年刊）》2013年第1期87-94,共8页Journal of Nanjing University(Mathematical Biquarterly)

摘　　要：利用两层线性规划的全局最优解可在其约束域的极点上达到这一性质,通过对问题可行解集合的结构进行探讨,将两层线性规划转化为带有反凸约束的线性规划,建立了一个新的全局解算法,证明了算法能收敛到问题的全局解,并通过一个算例说明了算法的求解过程.Based on the result that a global optimal solution for two-level linear programs occurs at an extreme point of its constraint region, we discusse the structural feature of its feasible region and converte the two-level linear program to linear programming with a reverse convex constraint, then build a new algorithm for the global solution, prove that the method converges to the global solution of the program. Finally, a simple example is given to illustrate the p^ovesa of the algorithm for the solution.

关 键 词：两层线性规划 全局最优解 反凸约束 极点 割平面 

分 类 号：O221.1[理学—运筹学与控制论]