无约束非线性l_p问题的区间极大熵方法  

Interval maximum entropy method for unconstrained nonlinear lpproblem

作　　者：赵建强[1] 李苏北[1] 陈必科 杨静[1] 

机构地区：[1]徐州工程学院数学与物理科学学院,江苏徐州221111

出　　处：《计算机应用研究》2014年第10期2974-2976,共3页Application Research of Computers

基　　金：国家自然科学基金青年项目(11001129);徐州工程学院校科研课题青年项目(XKY2010201)

摘　　要：针对信号处理、系统识别等领域中涉及到的无约束非线性lp问题,为减小由于二进制编码的舍入误差对该问题计算结果的影响,对求解该问题的极大熵方法进行了区间扩张。证明了区间扩张后的极大熵函数至少具有二阶收敛性,并设计了具有多项式时间复杂度的区间算法进行求解,举例进行了数值计算。数值计算结果显示,该区间算法可靠,计算结果与区间扩张前相比,结果更加精确。This paper discussed unconstrained nonlinear lp problem from signal processing, system identification and other ar- eas. In order to reduce the affection of the binary coded rounding errors on the calculation results of the problem, it introduced interval extension into the maximum entropy method. The new maximum entropy function improved by interval extension had at least second order convergence. It designed an interval algorithm with polynomial time complexity to solve some examples for the numerical solutions. The numerical results show that this new interval algorithm is reliable and the numerical solution using interval extension are more accurate.

关 键 词：非线性lp问题 极大熵 区间扩张 区间算法 

分 类 号：O221.2[理学—运筹学与控制论]