Jan 06, 2016
7/17/2019 LECTURE 1 of Value engineering and operations
1/29
Challenge the future
DelftUniversity ofTechnology
Value Engineering and OperationsOptimization (AE4441)
7/17/2019 LECTURE 1 of Value engineering and operations
2/29
2
Overview of the course
Wednesdays and Thursdays: Lectures/studio classroom
sessions on optimization methods!ntroduction to "perations #esearch$ %y &rederic' ()
*illier and +erald ,) Lie%erman- ninth edition
&inished .ith .ritten eam in 0rst period 1.ee' 34- notyet scheduled5) 6"T open3%oo')
7eriod
7eriod 2
T.o assignments &ocus on value engineering approach
(olve a value engineering pro%lem .ith optimization methods
Lectures on value engineering
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
3/29
8
&inal grade:
9am holds for 4 ; of the 0nal grade
9ach of the assignments holds for 24 ; of the 0nal grade
< grade = >)> for each of the elements is re?uired to pass the
course
+rades are valid for one year only
Overview of the course, continued
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
4/29
@
How to buy the book
The %oo' can %e %ought via the A(A
Use needs to %e made of the ne. A(A .e%shop) Bou can0nd this .e%shop under the oo' section at the .e%site 1
...)vsv)tudelft)nl5) Bou can pay via iDeal) The %oo' .ill
%e delivered at the desired address)
!ntroduction
http://www.vsv.tudelft.nl/http://www.vsv.tudelft.nl/7/17/2019 LECTURE 1 of Value engineering and operations
5/29
>
Overview of todays lecture
stpart Chapters -2-8)38)@
2ndpart
7/17/2019 LECTURE 1 of Value engineering and operations
6/29
Chapter 1: Introduction
The 0eld of "perations #esearch 1"#5 started in World War !!
This research on operations.as after.ards introduced in otherorganizations
"perations #esearch aims at determining optimal .ays to conduct
activities in an organization
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
7/29
Chapter 2: Overview of the OperationsResearch Modeling Approach
) De0ning the pro%lem and gathering data
2) Construct a mathematical model to represent the pro%lem
ndecision varia%les: the un'no.ns
"%Eective function: measure of performance Constraints: restrictions on decision varia%les
7arameters: constraint and o%Eective function constants
8) Develop a computer3%ased procedure for deriving solutions
to the pro%lem from the model
@) Test the model and re0ne if needed 1model validation5
>) 7repare for ongoing application of the model 1decisionsupport system5
) !mplement
Typical phases in an "# study:
6ote: these steps also hold for other0elds) !t holds for optimization approachesin general)
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
8/29
F
Chapter 2: Overview of the OperationsResearch Modeling Approach
) De0ning the pro%lem and gathering data
2) Construct a mathematical model to represent the pro%lem
ndecision varia%les: the un'no.ns
"%Eective function: measure of performance Constraints: restrictions on decision varia%les
7arameters: constraint and o%Eective function constants
8) Develop a computer3%ased procedure for deriving solutions
to the pro%lem from the model
@) Test the model and re0ne if needed 1model validation5
>) 7repare for ongoing application of the model 1decisionsupport system5
) !mplement
Typical phases in an "# study:
6ote: these steps also hold for other0elds) !t holds for optimization approachesin general)
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
9/29
G
ExampleResource optimization for operations
and maintenance of offshore wind farms5 7ro%lem: Large3scale roll3out of oHshore .ind energy
re?uires that its cost of electricity should %e reduced)
7otential for cost reduction in the 0eld of operations and
maintenance)
25The largest cost drivers are the vessels and technicians)
The follo.ing decision varia%les are identi0ed:
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
10/29
4
Iaimize
.here the availa%ility of the .ind3farm
days of preventive maintenance
costs 1salaries and e?uipment5
such that
minimum technician team size on a ("A
minimum technician team size on a CTA
- and are all integer values
)()()( DP CfCfAfY =
),,,,,( 654321 xxxxxxfA =
),,( 531 xxxfCP =
),,,,,(654321
xxxxxxfCD
=
%95A
qPMPM Re=
11 x
12 x
13 x
14
x 5x6x
654321 ,,,,, xxxxxx
#esulting mathematical model:
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
11/29
85 #esult o%tained %y computer3%ased procedure for
deriving solutions to the pro%lem from the model:
@5 Testing on eisting oHshore .indmill par's>5 Decision support tools for ongoing application 1daily
operations and ne. oHshore .indmill farms5
5 9nsure use of the tools
!ntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
12/29
2
Linear ProgrammingIntroduction
7/17/2019 LECTURE 1 of Value engineering and operations
13/29
8
Linear ProgrammingIntroduction, continued
Iost common application of linear programming:
*o. to allocate limited resources among competing activities in
a %est possi%le .ay
Chapter 8
Constraints Decision varia%les
"%Eective
function
7/17/2019 LECTURE 1 of Value engineering and operations
14/29
@
An example problemDefine the problem
The WB6D"# +L
7/17/2019 LECTURE 1 of Value engineering and operations
15/29
>
An example problemDefine the problem
ecause of declining earnings- it has %een decided to revamp the companyKs
product line) Unpro0ta%le products are %eing discontinued) This releases
production capacity to launch t.o ne. products having large sales potential
7roduct : < glass door .ith aluminum framing
7roduct 2: < dou%le3hung .ood3framed .indo.
7roduct re?uires some of the production capacity in 7lants and 8- %ut none in
7lant 2) 7roduct 2 needs only 7lants 2 and 8
The mar'eting division has concluded that the company could sell as much of
either product as could %e produced %y these plants) *o.ever- %ecause %oth
products .ould %e competing for the same production capacity in 7lant 8- it is not
clear .hich mi of the t.o products .ould %e most pro0ta%le)
Chapter 8
7/17/2019 LECTURE 1 of Value engineering and operations
16/29
An example problemGather data
&irst values for the parameters needed
Chapter 8
Plant
Production time perbatch (Hours)
Product
1 2
Production timeavailable per week
(Hours)
1
2
3
1 0
0 2
3 2
4
12
18
Proft per batch 3000 !000
6um%er of hours of production time availa%le per .ee' in each plant for the ne. products
6um%er of hours of production time used in each plant for producing the ne. products
7ro0t for each of the ne. products
7/17/2019 LECTURE 1 of Value engineering and operations
17/29
Decision varia%les:
x: num%er of %atches of product produced per .ee'
x2: num%er of %atches of product 2 produced per .ee'
The o%Eective function
Z 8x M >x2 1total pro0t per .ee' in thousands of dollars5
The aim is to maimizeZ
Constraints
x @
2x22
8xM 2x2F
and
x4
x24
Chapter 8
An example problemIts linear programming model
6um%er of hours of production time availa%le per .ee' in plant
6um%er of hours of production time availa%le per .ee' in plant 2
6um%er of hours of production time availa%le per .ee' in plant 8
7/17/2019 LECTURE 1 of Value engineering and operations
18/29
F
Iaimize Z 8x M >x2
(u%Eect to
x @2x228xM 2x2F
x4
x24
Chapter 8
An example problemIts linear programming model
7/17/2019 LECTURE 1 of Value engineering and operations
19/29
G
An example problemDerive solutions
IaimizeZ 8x M >x2
x@
(u%Eectto
2x22
8xM 2x2F
Chapter 8
7/17/2019 LECTURE 1 of Value engineering and operations
20/29
24
An example problemFind optimal solution
Chapter 8
Z 8xM
>x2
&easi%leregion
Z 24
Z 84Z 8
7/17/2019 LECTURE 1 of Value engineering and operations
21/29
2
&or this situation a graphical approach suJces) The optimal solution is:
x 2x2
Z 8
- i)e) maimal pro0t .ithin the constraints is o%tained for a mi of 2
%atches of product per .ee'- and %atches of product 2 per .ee')
The resulting total pro0t is N8444 per .ee'
An example problemOptimal solution
7roduct : glass door .ith aluminum frami7roduct 2: Dou%le3hung .ood3framed .ind
Chapter 8
7/17/2019 LECTURE 1 of Value engineering and operations
22/29
22
Iaimize Z cxM c2x2M O) M cnxn
(u%Eect to
axM a2x2M O M anxnb
a2xM a22x2M O M a2nxnb2
amxM am2x2M O M amnxnbm
and
x4-x24- O-xn4
Chapter 8
Linear ProgrammingThe mathematical model, its general form
&unctionalconstraints
6onnegativity
constraints
6ote: other forms also possi%le
"%Eective function
7/17/2019 LECTURE 1 of Value engineering and operations
23/29
28Chapter 8
Linear ProgrammingThe mathematical model, its solutions
&easi%le solution:
7/17/2019 LECTURE 1 of Value engineering and operations
24/29
2@Chapter 8
Linear ProgrammingThe mathematical model, its solutions
The importance of C7& solutions:
7/17/2019 LECTURE 1 of Value engineering and operations
25/29
2>Chapter 8
Linear ProgrammingThe assumptions, proportionality
7roportionality:
The contri%ution of each activityjto the value of theo%Eective functionZis proportional to the level of activityxE1cExE5
andThe contri%ution of activityjto the left3hand3side of each
functional constraint is proportional to the level ofactivityxE1aiExE5
These situations
prevent a1straightfor.ard5 linearprogramming approachto solve for optimalsolutions for thedecision varia%les)
7roportionality violated
7/17/2019 LECTURE 1 of Value engineering and operations
26/29
2Chapter 8
Linear ProgrammingThe assumptions, additivity
x23xx2
Z 8x M >x2Mxx2
&unctions that do not satisfy this re?uirement- e)g):
7/17/2019 LECTURE 1 of Value engineering and operations
27/29
2Chapter 8
Linear ProgrammingThe assumptions, divisibility
Divisi%ility:
Decision varia%les in a linear programming model areallo.ed to have any values- including non3integer values-that satisfy the functional and non3negativity constraints
7/17/2019 LECTURE 1 of Value engineering and operations
28/29
2FChapter 8
Summary
7ro%lems dealing .ith assigning scarce resources to aseries of activities in an optimal .ay- can often %e solvedthrough linear programming
*ereto a mathematical model is esta%lishedThe o%Eective function rePects the measure of
performanceThe constraints rePect the limited resources
< search needs to %e carried to 0nd those levels ofactivities that maimize the o%Eective function and still
are .ithin the constraints
Chapter 8 considers a graphical approach) This is notfeasi%le .hen more than three levels of activity need to%e determined
7/17/2019 LECTURE 1 of Value engineering and operations
29/29
2GChapter 8
Summary, continued
The feasi%le region is the collection of all possi%lesolutions 1levels of activity5 that satisfy all constraints
Corner3point feasi%le 1C7&5 solutions are located at acorner of the feasi%le region
These C7& solutions are of interest %ecause the %estC7& solution must %e an optimal solution) !f the pro%lemhas multiple optimal solutions at least t.o must %e a C7&solution
The assumption of a linear o%Eective function su%Eect tolinear constraints is often allo.ed for real life pro%lems