• No results found

Duality and algorithmic aspects of certain integer nonlinear programming problems

N/A
N/A
Protected

Academic year: 2022

Share "Duality and algorithmic aspects of certain integer nonlinear programming problems"

Copied!
6
0
0

Loading.... (view fulltext now)

Full text

(1)

DUALITY AND ALGORITHMIC ASPECTS OF CERTAIN INTGER NONLINEAR PROGRAMMING PROBLEMS

by

M. CHANDRAMOHAN

Department of Mathematics

Indian Institute of

Technology,

Delhi

Submitted

in fulfilment of the requirements of the degree of DOCTOR OF PHILOSOPHY

to the

Indian

Institute of

Technology, Delhi December, 1978

(2)

ACKNOWLEDGEMENTS

It gives me great pleasuru to place on record my deep sense of gratitude to my supervisor Dr. Suresh Chandra, Deptt.

of Mathematics, Indian Institute of Technology, Delhi, for the invaluable help I had received from him throughout my research work here. But for his keen interest he had evinced in the progress of my studies, this would not have been possible.

My thanks are due to Professor O.P. Bhutani, Head cf the Department and Professor N.S. Kambo, Department of Mathema- tics, Indian Institute of Technology, Delhi for their encourage- ment and interest in my work.

I thank the Director, I.I.T. Delhi, for providing me all the facilities for research work and also acknowledge the finan- cial assistance I had received from the Ministry

•f

Education, Government of India, in the form of Q.I.P. Scholarship for three years. My profound thanks go to Principal, Calicut Regional Engineering College, Calicut, for sponsoring me under Q.I.P. for Ph.D. studies.

The kindness of Mathematical community in sending me reprints and preprints has been imMense which I gratefully acknowledge.

My thanks are also due to my friends and colleagues and in particular to Sqn. Ldr. L.S. Vaidyanathan and Mr. K.J. George for all the help I had received from them.

I owe a debt of gratitude to my wife whose sacrifice and encouragement have always been a great source of strength to me.

I would also like to express my deep sense of indebtedness to my parents, brothers and sisters for their tremendous patience, endurance and affection.

Finally, Mr. Ranjit Kumar and Miss Neelam Dhody deserve praise for their expert typing of this thesis.

M. atiAnRAmOHAN

(3)

ABSTRACT

This thesis entitled "Duality and Algorithmic Aspects of Certain Integer Nonlinear Programming Problems" is divided into six chapters, the contents of which are given below.

The first chapter is general intrioduction which con- tains preliminaries, review of related work and a summary of the thesis.

The second chapter is on duality in mixed integer linear and nonlinear fractional programming and integer nondifferentiable programming. This also includes a method for solving integer nondifferentiable programs based on the above duality theory.

The third chapter consists of a study of variable

transformation methods for solving integer fractional programs.

In this, the equivalence of integer fractional programs with certain related problems whose objective functions are no longer fractionaIl are established. The study also includes development of branch and bound methods for solving integer linear and nonlinear fractional programs based on these equivalence relations.

The fourth chapter is on parametric methods for solving fractional programs. Here, solvability of mixed integer linear fractional programs through a finite number

(4)

of mixed integer linear programs is established. It also contains procedures for solving integer linear fractional programs. Parametric methods for solving integer nonlinear fractional programs are also proposed.

The fifth chapter consists of a method for enforcing additional constraints to linear fractional programs and its applications in solving integer linear fractional pro- grams by using cutting plane and branch and bound methods.

The sixth and the last chapter of the thesis consists of a finite all integer method for solving integer convex programs whose problem functions are polynomials with

rational coefficients. The applicability of this in solving certain integer nonlinear fractional programs is also

pointed out.

(5)

CONTENTS

Page

CHAPTER I INTRODUCTION 1

1.1 Preliminaries 4

1.2 Review of Related Work 9 1.3 Summary of the Thesis 23 CHAPTER II DUALITY IN INTEGER NONCONVEX AND

NONDIFFERENTIABLE PROGRAMMING 30.

2.1 Mixed Integer Linear Fractional

Programming 32

2.2 Mixed Integer Nonlinear Fractional

Programming 1+0

2.3 Integer Nondifferentialle

Programming 50

2.4 Algorithmic Aspects Based on

Duality 66

CHAPTER III VARIABER TRANSFORMATION METHODS FOR

INTEGER ERACTIONAL PROGRAMS 72 3.1 The Equivalent Problem for an

Integer Linear Fractional Program

(ILFP) 75

3.2 A Branch and Bound Method for Integer Linear Fractional

Programs (ILFPs) 79

3.3

Integer Nonlinear Fractional

Programs 95

3.4 Linear Fractional Complementary

Programming Problems 109 3.5 Sensitivity Analysis in Integer

Linear Fractional Programming

(ILFP) 112

CHAPTER IV PARLEIRIG METHODS FOR INTEGER

FRACTIONAL PROGRAMS 11 5

4.1 Parametric Method for Mixed Integer Linear Fractional

Programs 116

(6)

Page CHAPTER IV (Continued)

4.2 Methods of Solution 122 4.3 Generalization 131+

CHAPTER V A METHOD OF ENFORCING ADDITIONAL CONSTRAINTS TO LINEAR FRACTIONAL PROGRAMS AND ITS USES

IN SOLVING INTF.GER LINEAR FRACTIONAL PROGRAMS 140 5.1 Development of the Method 142 5.2 Applications in Solving Integer

Linear Fractional Programs (ILFPs) 153 CHAPTER VI AN ALL INILGER METHOD FOR INMGER COJVEX

PROGRAMS 168

6.1 Development of the Method 169 6.2 Finiteness of the Algorithm 175 6.3 Relaxing the Restrictions 178 6.4 Extension to Integer Nonlinear

Fractional Programs (INLFPs) 183

REliERENCES 185

References

Related documents

It was found that solving RWA problem using genetic al- gorithm minimizes the congestion but total route length of the lightpaths increased compared to shortest path

In this

The main difference between the two is that our objective function is now a quadratic function (the constraints are still linear).Because of its many applications, quadratic

Also many powerful methods, for example, exponential function method [7,8], (G /G) -expan- sion method [9,10], first integral method [11,12], sub- equation method [13,14],

By using a stabilization theorem proposed newly for a class of nonlinear systems, linear feedback controller is designed to stabilize the fractional-order system and the

So we have a logarithmic enhancement for the relative ratio of the susceptibility of the nonlinear fractional relaxation with respect to nonlinear integer relaxation. The first

Employing this method the fractional volumes of basins of attraction have been computed for three cases; firstly for the linear programming rule with radii of maximal

Inverse of a Matrix and its Application in Solving System of Linear Equations Paper No: 14Statistical Applications in Environmental Sciences.. Module: 24 Inverse of a Matrix and