Files in this item



application/pdf8203390.pdf (3MB)Restricted to U of Illinois
(no description provided)PDF


Title:Integer Programming Heuristics for Large Capital Budgeting Problems
Author(s):Agori-Iwe, Kesiena Oghenemado
Department / Program:Agricultural Economics
Discipline:Agricultural Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Economics, Agricultural
Abstract:The subject matter of this thesis was the comparison of five integer algorithms suitable for solving large capital budgeting problems such as those characterized by the national development plans in developing nations. In order to obtain statistically meaningful results, we designed and carried out some nonparametric randomized complete block experiments using real world data from the 1968/69 Nigerian national development plan.
Our results indicate that there is a negative correlation between the solution time and the relative accuracy of the algorithms. The heuristic or approximative algorithms were generally much faster in obtaining results than the exact algorithm. Although the fastest of these algorithms, the Effective Gradient Method, also had the largest relative error, none of the heuristic algorithms returned a relative error of more than 7.0 per cent for the 10 subsets of 100 variables drawn from 1000 variables in the experiments.
A new heuristic algorithm, the Pseudo Reduction Method, which was developed in this study to exploit the special structure of the bounded variable knapsack problem which was exhibited in the development plan, returned more accurate and faster solutions when compared with the Kochenberger/McCarl/Wyman algorithm, another method which was originally designed to solve bounded or general integer variable problems.
Issue Date:1981
Description:116 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1981.
Other Identifier(s):(UMI)AAI8203390
Date Available in IDEALS:2014-12-13
Date Deposited:1981

This item appears in the following Collection(s)

Item Statistics