A QUADRATICALLY CONVERGENT GLOBAL ALGORITHM FOR THE LINEARLY-CONSTRAINED MINIMUM CROSS-ENTROPY PROBLEM

被引：4

作者：

FANG, SC

TSAO, HSJ

机构：

[1] N CAROLINA STATE UNIV,OPERAT RES PROGRAM,RALEIGH,NC 27695

[2] N CAROLINA STATE UNIV,DEPT IND ENGN,RALEIGH,NC 27695

[3] UNIV CALIF BERKELEY,INST TRANSPORTAT STUDIES,BERKELEY,CA 94720

来源：

EUROPEAN JOURNAL OF OPERATIONAL RESEARCH | 1994年 / 79卷 / 02期

关键词：

ENTROPY OPTIMIZATION; DUALITY; UNCONSTRAINED OPTIMIZATION; CURVED-SEARCH ALGORITHM;

D O I：

10.1016/0377-2217(94)90365-4

中图分类号：

C93 [管理学];

学科分类号：

12 ; 1201 ; 1202 ; 120202 ;

摘要：

In this paper, we propose a curved-search algorithm for solving the cross-entropy minimization problem with linear equality constrains. The proposed algorithm converges globally to a dual optimal solution with a quadratic rate of convergence. A dual-to-primal conversion formula is provided. We also analyze the computational effort required for the algorithm and report our computational experience.

引用

页码：369 / 378

页数：10

共 38 条

[1] DISCRETE CHOICE THEORY, INFORMATION-THEORY AND THE MULTINOMIAL LOGIT AND GRAVITY MODELS [J].