摘要
目的讨论变量有界线性规划问题的熵函数解法。方法采用Lagrangian对偶把该问题处理为一个对偶的低维无约束不可微凸规划,据此提出了变量有界线性规划问题的可微极大熵函数。结果提出的熵函数方法可以避免数值计算的溢出。结论所采用的熵函数可避免数值的溢出,数字结果表明方法是有效的。
Aim To give a maximum entropy method for linear programming with bounded variables. Methods The Lagrangian dual is used to deal with the problem and an unconstraint nondifferentiable optimization problem is obtained. Results An entropy function is given to solve the nondifferentiable problem, and this entropy function can avoid arithmetic overflow. Conclusion The used enfropy function avoids the numerical oueflow. Numerical examples are given to show the efficiency of the method.
出处
《西北大学学报(自然科学版)》
CAS
CSCD
北大核心
2005年第5期507-510,共4页
Journal of Northwest University(Natural Science Edition)
基金
国家自然科学基金资助项目(60072044)