Scribd
Upload
211 views105 pages

Hock Schittkowski Collection

This appendix provides a detailed description of test problems from an out-of-print monograph on nonlinear programming. Each problem is classified using a scheme that characterizes the mathematical structure of the objective function and constraints. The classification includes information about the regularity of the problem, whether an exact solution is known, the order of derivatives implemented, and a serial number. For some problems, program fragments or additional data are provided where analytical expressions alone are insufficient to describe the objective and constraint functions.

Uploaded by

Michel Alba
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
211 views105 pages

Hock Schittkowski Collection

This appendix provides a detailed description of test problems from an out-of-print monograph on nonlinear programming. Each problem is classified using a scheme that characterizes the mathematical structure of the objective function and constraints. The classification includes information about the regularity of the problem, whether an exact solution is known, the order of derivatives implemented, and a serial number. For some problems, program fragments or additional data are provided where analytical expressions alone are insufficient to describe the objective and constraint functions.

Uploaded by

Michel Alba
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 105

APPENDIX: Test Problems of the First Collection

Purpose of this appendix is to list a detailed description of all test problems published
in the monograph [4], which is out of print. We proceed from the nonlinear program (1)
and list the following data of an example:
PROBLEM: test problem number
CLASSIFICATION: classication number in the form OCD-Kr-s ac-
cording to the scheme given below
NUMBER OF VARIABLES: number of variables n
NUMBER OF CONSTRAINTS: number of inequality constraints, m1 , number of
equality constraints, i.e., m m1 , and number of
variable bounds of variables, b
OBJECTIVE FUNCTION: analytical expressions for objective function f (x)
CONSTRAINTS: analytical expressions for constraints gj (x), j =
1, . . . , m
START: starting values for variables, x0 , and corresponding
objective function value, f (x0 ), together with an
information whether x0 is feasible or not
SOLUTION: information about optimal solution x , i.e.,
- objective function value f (x )
- constraint violation, r(x )
- norm of gradient of Lagrange function
- number of active constraints,
- active constraints, I(x )
- degree of degeneracy, umax /umin
- condition number of projected Hessian of La-
grange function, max /min

The general form of the classication scheme is

OCD-Kr-s

with
O - objective function
C - constraints
D - regularity
K - information about solution, i.e., whether an exact solution is known or not
r - order of partial derivatives
s - serial number within a class

The purpose of the classication scheme is to characterize the mathematical structure


of objective function and constraints, and to give more information about the implemen-
tation and the solution. A problem is called a regular one, if rst and second derivatives

17
exist in the feasible region for all problem functions, otherwise an irregular one. The
subsequent abbreviations are used:
class key description
O C constant function
L linear function
Q quadratic function
S sum of squares
P generalized polynomial function
G general function
C U unconstrained problem
B only upper and lower bounds
L linear functions
Q quadratic functions
P generalized polynomial functions
G general functions
D R regular problem
I irregular problem
K T exact solution known (theoretical problem)
P exact solution not known (practical problem)
r 0 derivatives not implemented
1 rst derivatives implemented

For some test problems, we cannot describe objective or constraint functions just by a
few analytical expressions. In these cases, program fragments are attached at the end of
this section together with more extensive information about a test problem, e.g., constant
data, starting or solution values.
The subsequent pages are xeroxed copies of the original publication.

18

You might also like