Upload
aviraj2006
View
219
Download
0
Embed Size (px)
Citation preview
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 1/19
D Nagesh Kumar, IISc Optimization Methods: M1L31
Introduction and Basic Concepts
(iii) Classificationof Optimization
Problems
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 2/19
D Nagesh Kumar, IISc Optimization Methods: M1L32
Introduction
Optimization problems can be classified based on
the type of constraints, nature of design variables,
physical structure of the problem, nature of theequations involved, deterministic nature of the
variables, permissible value of the design variables,
separability of the functions and number of objectivefunctions. These classifications are briefly
discussed in this lecture.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 3/19
D Nagesh Kumar, IISc Optimization Methods: M1L33
Classification based on existence of constraints.
Constrained optimization problems: which are
subject to one or more constraints.
Unconstrained optimization problems: in which
no constraints exist.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 4/19D Nagesh Kumar, IISc Optimization Methods: M1L34
Classification based on the nature of the designvariables
There are two broad categories of classificationwithin this classification
First category : the objective is to find a set of designparameters that make a prescribed function of theseparameters minimum or maximum subject to certain
constraints. – For example to find the minimum weight design of a strip
footing with two loads shown in the figure, subject to alimitation on the maximum settlement of the structure.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 5/19
The problem can be defined as follows
subject to the constraints
The length of the footing (l) the loads P1 and P2 , the distance between the loads areassumed to be constant and the required optimization is achieved by varying b and d.
Such problems are called parameter or static optimization problems.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 6/19
D Nagesh Kumar, IISc Optimization Methods: M1L36
Classification based on the nature of the designvariables (contd.)
Second category: the objective is to find a set ofdesign parameters, which are all continuous
functions of some other parameter, that minimizesan objective function subject to a set of constraints.
– For example, if the cross sectional dimensions of therectangular footing is allowed to vary along its length asshown in the following figure.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 7/19
The problem can be defined as follows
subject to the constraints
l
The length of the footing (l) the loads P1 and P2 , the distance between the loads areassumed to be constant and the required optimization is achieved by varying b and d.
Such problems are called trajectory or dynamic optimization problems.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 8/19
D Nagesh Kumar, IISc Optimization Methods: M1L38
Classification based on the physical structure of theproblem
Based on the physical structure, we can classifyoptimization problems are classified as optimal control and
non-optimal control problems.
(i) An optimal control (OC) problem is a mathematicalprogramming problem involving a number of stages, whereeach stage evolves from the preceding stage in aprescribed manner.
It is defined by two types of variables: the control or designvariables and state variables.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 9/19
The problem is to find a set of control or design variables such that thetotal objective function (also known as the performance index, PI) over
all stages is minimized subject to a set of constraints on the control andstate variables. An OC problem can be stated as follows:
Where x i is the i th control variable, y i is the i th state variable, and f i is thecontribution of the i th stage to the total objective function. gj, hk, and q i
are the functions of xj, yj ; x k, y k and x i and y i , respectively, and l is thetotal number of states. The control and state variables x i and y i can bevectors in some cases.
(ii) The problems which are not optimal control problems are
called non-optimal control problems.
subject to the constraints:
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 10/19
D Nagesh Kumar, IISc Optimization Methods: M1L310
Classification based on the nature of the equationsinvolved
Based on the nature of expressions for the objectivefunction and the constraints, optimization problems can
be classified as linear, nonlinear, geometric andquadratic programming problems.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 11/19
D Nagesh Kumar, IISc Optimization Methods: M1L311
Classification based on the nature of the equationsinvolved (contd.)
(i) Linear programming problem
If the objective function and all the constraints are linear functions ofthe design variables, the mathematical programming problem is calleda linear programming (LP) problem.
often stated in the standard form :subject to
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 12/19
D Nagesh Kumar, IISc Optimization Methods: M1L312
Classification based on the nature of theequations involved (contd.)
(ii) Nonlinear programming problem
If any of the functions among the objectives and constraintfunctions is nonlinear, the problem is called a nonlinearprogramming (NLP) problem this is the most general form of a
programming problem.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 13/19
D Nagesh Kumar, IISc Optimization Methods: M1L313
Classification based on the nature of the equationsinvolved (contd.)
(iii) Geometric programming problem
– A geometric programming (GMP) problem is one in which theobjective function and constraints are expressed as polynomials in X.
A polynomial with N terms can be expressed as
– Thus GMP problems can be expressed as follows: Find X which
minimizes : subject to:
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 14/19
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 15/19
D Nagesh Kumar, IISc Optimization Methods: M1L315
Classification based on the permissible values ofthe decision variables
Under this classification problems can be classified asinteger and real-valued programming problems
(i) Integer programming problem
If some or all of the design variables of an optimization problem arerestricted to take only integer (or discrete) values, the problem iscalled an integer programming problem.
(ii) Real-valued programming problem
A real-valued problem is that in which it is sought to minimize ormaximize a real function by systematically choosing the values of realvariables from within an allowed set. When the allowed set containsonly real values, it is called a real-valued programming problem.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 16/19
D Nagesh Kumar, IISc Optimization Methods: M1L316
Classification based on deterministic nature of thevariables
Under this classification, optimization problemscan be classified as deterministic and stochasticprogramming problems
(i) Deterministic programming problem • In this type of problems all the design variables are
deterministic.
(ii) Stochastic programming problem In this type of an optimization problem some or all the
parameters (design variables and/or pre-assigned parameters)are probabilistic (non deterministic or stochastic). For exampleestimates of life span of structures which have probabilisticinputs of the concrete strength and load capacity. A
deterministic value of the life-span is non-attainable.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 17/19
D Nagesh Kumar, IISc Optimization Methods: M1L317
Classification based on separability of the functions
Based on the separability of the objective and constraintfunctions optimization problems can be classified asseparable and non-separable programming problems
(i) Separable programming problems
In this type of a problem the objective function and the constraints areseparable. A function is said to be separable if it can be expressed asthe sum of n single-variable functions and separable programmingproblem can be expressed in standard form as :
subject to :
where b j is a constant.
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 18/19
D Nagesh Kumar, IISc Optimization Methods: M1L318
Classification based on the number of objectivefunctions
Under this classification objective functions can beclassified as single and multiobjective programmingproblems.
(i) Single-objective programming problem in which there is only a
single objective.(ii) Multi-objective programming problem
A multiobjective programming problem can be stated as follows:
where f 1, f 2 , . . . f k denote the objective functions to be minimizedsimultaneously.
subject to :
8/3/2019 M1L3slides
http://slidepdf.com/reader/full/m1l3slides 19/19
D Nagesh Kumar, IISc Optimization Methods: M1L319
Thank You