Alif Muhammad Naufal

13418043
Rachmani Aisha N
134184045
Farhan Hakim
13418055

Multiple Constraint:
Moh. Abi Hamid
13418057
Amarta Dhimam Prasaja

Linear Programming 13418068

Aditya Putra Budiyanto
13418070
Chapter 14 Felicia Collins Willim
13418080
 CONTENTS
1 Constrained Optimization

2 A Product Mix Example

3 A Linear Programming Model

4 Solution by Computer

Effect of Forcing Production of

5 Luxury

6 Pineapple Delight Case Study

7 A Transportation Problem

2
Constrained Optimization
14.1
Optimization with Multiple 4
Constraints
 In a multiple constraint decision problem, some of the constraints
may be binding on the optimal solution, while others will not be.
 If the constraints are all in the form of scarce resources, a binding
constraint implies that all of the resource is used up, it means the
total amount of the resource consumed is equal to the amount
available.
 A constraint that is not binding means that some of the resource
remains unused, it means the constraint has slack.
 The optimal constrained solution can never be better than the
optimal unconstrained solution.
Optimization with Multiple 5
Constraints
 If we knew which constraints are binding on the decision variables
at the optimal solution, we could simply discard all other non-
binding constraints. These will, by definition, have no effect on the
optimal constrained values of the decision variables.
 Unfortunately, there is no simple way of identifying which constraints
are binding and which constraints have slack.
Optimization with Multiple 6
Constraints
Contoh Soal Variabel keputusan:
𝑥1 = 13 dan 𝑥2 = 2
Maksimasi:
𝑍 = 3𝑥1 + 2𝑥2 Fungsi objektif:
𝑍 = 43
Fungsi pembatas:
𝑥1 + 𝑥2 ≤ 15
Constraint has
2𝑥1 + 𝑥2 ≤ 28 slack
𝑥1 + 2𝑥2 ≤ 20
𝑥1 , 𝑥2 ≥ 0
 7

An Algorithmic
Analogy
Local and Global Optima 8

Is this the highest

point on the island?
Local and Global Optima 9

 If the island has only a single peak and the ground rises from the
shore steadily towards this single peak from every feasible direction,
even along the cliff edges, then you can be sure that you have
reached the highest point even in pitch darkness. In mathematic
context, we say that the surface of the island is well-behaved.
 In fact, many islands have many high points with peaks of various
heights and possibly also some plateaux and valleys part-way up,
we say that the surface of the island is not well-behaved.
 Then, there is no guarantee that this algorithm will lead you to the
highest point on the island.
Local and Global Optima 10

 If the mathematical form is well-behaved, an algorithm which at

each iteration improves the value of the objective function, while
remaining feasible, will ultimately find the optimal solution.
 If a problem doesn’t have a well-behaved mathematical form,
then no such guarantee can be given. You might have found a
local optimum or the global optimum.
 In linear programming, our focus is to solve the well-behaved form.
A Product Mix
Example

14.2
A Situation Summary

Sebuah Perusahaan furniture memproduksi stasiun kerja computer yang ergonomis.

Stasiun kerja tersebut terdiri dari:
• Tubular frame yang disesuaikan dengan keyboard
• Monitor
• Komputer
• Printer
• Dan Aksesoris lainnya

Terdapat 3 model yang dibuat yaitu:

• Model Basic
Dijual di gerai ritel
• Model Standard
perusahaan itu
• Model Luxury
sendiri

12
Dalam pembuatan stasiun kerja tersebut • Operasi pengelasan, pembuatan rak, dan
terdapat 5 operasi yang diperlukan yaitu:
perakitan semuanya membutuhkan waktu
1. Memotong tubular frame → 1 pekerja setup harian, karena itu terdapat waktu
2. Mengelas frame → 2 pekerja produksi yang hilang.
3. Mengecat frame
• Fasilitas pengecatan dan perawatan semprot
4. Membuat rak → 3 pekerja memiliki kapasitas 32 frame per hari apapun
5. Merakit frame, rak, dan roller → 2 pekerja jenis modelnya

13
 Hard Soft
constrain constrain Berasal dari
Berasal dari teknis
kebijakan
saat produksi dan
pengelola yang
fasilitas.
disengaja.

Jumlah model
Kapasitas masing- standard
masing operasi sedikitnya 1/3 dari
jumlah model basic

14
 Bagaimana bauran
produk yang efektif
untuk meningkatkan
laba kotor dengan
batasan-batasan
tersebut yaaa…?

Manajemen 15
 Minutes of Welding
Minutes of
Cutting time
cutting time
Number of each welding time
available / time / unit available /
day / unit type of workstation produced day
produced
produced/day

Total Status of
Status of
Total welding welding
cutting
cutting time used time
time
time used / day constrain
constrain
/ day
Revenue
Gross and
Total profit material
number cost / unit
Status of of unit / Basic/
painting day standard
constrain Total output
shelving Status of Total Status of relationship
time used assembly required
shelving assembly
/ day constrain time used constrain
/ day
Painting
capacity / Minutes of
time for Status
day Shelving
shelving /
capacity /
Minutes of Assembly basic to
unit assembly capacity /
day time / unit day
standard
output
16
14.3

A Linear
Programming Model
 • A decision choice consists of the number of
units produced per day for each product.
Rather than denote the variables by letter
symbols, such as 𝑥1 , 𝑥2 , and 𝑥3 , it is more
helpful to use mnemonic names.
• The modelling system M assumes that all
Decision decision variables may take on any value,
integer or fractional that is non-negative.
Variables • If the variables can only assume integer
values, the the problem has to be
formulated as an integer linear
programming model.

18
 Objective Function for Product Mix Problem

Material
Revenue Difference Number of Gross Profit per
Type Cost per
per Unit per Unit Units/Day Day
Unit
Basic $143 $43 $100 BASIC 100 BASIC
Standard $180 $60 $120 STANDARD 120 STANDARD
Luxury $246 $86 $160 LUXURY 160 LUXURY

Total/day: Gross profit = 100 BASIC + 120 STANDARD + 160 LUXURY

19
 Restrictions on Decision Variable Choice

DECISION VARIABLE
BASIC STANDARD LUXURY sign RHS
• To be a feasible solution to the product 100 120 160 MAX
Objective : GROSSPROF
mix problem, the three decision variables
Constraints CUTTING 16 12 20 <= 480
have to be non-negative and satisfy all six
constraints. WELDING 25 22,5 36 <= 900
PAINTING 1 1 1 <= 32
• The optimal solution to the problem is
the feasible solution that maximizes the SHELVING 36 50 80 <= 1380

objective function. ASSEMBLY 22,5 25 40 <= 900

BAS/STD MIX -0,333333333 1 0 >= 0

20
Solution by
Computer

14.4
Solution by Computer

Untuk mengatasi keterbatasan jika menghitung secara manual untuk perhitungan Linear Programming
berukuran besar, kita dapat menghitung dengan bantuan software. Hasil dapat diperoleh dalam waktu
yang singkat

Software komputer yang dapat menyelesaikan permasalahan LP:

• Spreadsheet packages dengan fungsi optimizer atau solver dan add-on mathematical programming
routines (semakin tidak akurat jika ukuran bertambah)
• Untuk LP dengan 100 atau lebih variabel dan/atau constraint, digunakan LP software (CPLEX,
XPRESS, OSL) atau mathematical optimization software (AMPL, AIMMS, GAMS, MPL)
• LINDO sering digunakan sebagai educational LP package, tetapi mampu untuk mengatasi
permasalahan dengan ukuran yang wajar

22
 Decision variables
BASIC STANDARD LUXURY sum sign RHS
Changing
cells

Excel Solver Objective:

GROSSPROF
100 120 160 max

CUTTING 16 12 20 ≤ 480
WELDING 25 22.5 36 ≤ 900
PAINTING 1 1 1 ≤ 32
Constraints
SHELVING 36 50 80 ≤ 1380
ASSEMBLY 22.5 25 40 ≤ 900
BAS/STD MIX -0.3333 1 ≥ 0

23
24
Menambah constraint

25
 Decision variables
BASIC STANDARD LUXURY sum sign RHS
Changing
15.71429 16.28571 0
cells

Hasil Objective:
GROSSPROF
100 120 160 3525.714 max

CUTTING 16 12 20 446.8571 ≤ 480

WELDING 25 22.5 36 759.2857 ≤ 900
PAINTING 1 1 1 32 ≤ 32
Constraints
SHELVING 36 50 80 1380 ≤ 1380
ASSEMBLY 22.5 25 40 760.7143 ≤ 900
BAS/STD MIX -0.3333 1 11.04814 ≥ 0

26
Answer
Report

27
Binding or Decision variables

Not Binding Changing

BASIC

15.71429
STANDARD LUXURY

16.28571 0
sum sign RHS

cells
• Status suatu constraint Objective:
100 120 160 3525.714 max
GROSSPROF
dapat dikategorikan
CUTTING 16 12 20 446.8571 ≤ 480
menjadi 'binding' atau WELDING 25 22.5 36 759.2857 ≤ 900
'not binding' Constraints
PAINTING
SHELVING
1
36
1
50
1
80
32
1380
≤
≤
32
1380
berdasarkan slack ASSEMBLY 22.5 25 40 760.7143 ≤ 900
BAS/STD MIX -0.3333 1 11.04814 ≥ 0
• Slack adalah selisih
antara RHS dan level
constraint yang
dicapai

28
Sensitivity
Report

29
Reduced Cost
• Akan bernilai 0 untuk variabel keputusan yang bernilai positif
• Untuk variabel keputusan yang bernilai 0, reduced cost adalah
perubahan/selisih nilai fungsi objektif akibat menetapkan nilai
variabel keputusan menjadi 1. (Akibatnya, nilai variabel keputusan
lainnya juga dapat ikut berubah akibat ini)
• Untuk kasus maksimasi, reduced cost bernilai negatif atau nol yang
menandakan kemungkinan penurunan (possible decrease) fungsi
objektif akibat penambahan nilai variabel keputusan dari 0 menjadi 1
• Untuk kasus minimasi, reduced cost bernilai positif atau nol yang
menandakan kemungkinan kenaikan (possible increase) fungsi
objektif akibat penambahan nilai variabel keputusan dari 0 menjadi 1
30
 Decision variables
BASIC STANDARD LUXURY sum sign RHS
Changing 13.1428571
17.85714 1
cells 4
Objective:
100 120 160 3525.857 max
GROSSPROF
CUTTING 16 12 20 463.4286 ≤ 480
WELDING 25 22.5 36 778.1429 ≤ 900
PAINTING 1 1 1 32 ≤ 32
Constraints
SHELVING 36 50 80 1380 ≤ 1380
ASSEMBLY 22.5 25 40 770.3571 ≤ 900
BAS/STD MIX -0.3333 1 7.191071 ≥ 0

Reduced Cost
= 3525.857-3525.714
= -2.85714

31
 Ini berarti mengubah nilai
Bagaimana kalau variabel keputusan tidak akan
mengakibatkan perubahan
reduced cost = 0
pada nilai fungsi objektif.
ketika nilai variabel Artinya, permasalahan LP ini
keputusan optimal = 0? memiliki banyak alternatif
solusi.

32
 Jika kita menambahkan negatif dari reduced cost (-2.85714) ke koefisien fungsi
objektif, ini akan mengakibatkan nilai variabel keputusannya yang semula 0
menjadi suatu nilai positif
Decision variables
STANDAR
BASIC LUXURY sum sign RHS
D
Changing
15.71429 16.28571 0
cells
Objective:
AFTER
GROSSPRO 100 120 160 3525.714 max
F
CUTTING 16 12 20 446.8571 ≤ 480 Decision variables
WELDING 25 22.5 36 759.2857 ≤ 900 BASIC STANDARD LUXURY sum sign RHS
PAINTING 1 1 1 32 ≤ 32 Changing
Constraints 20 10 2
SHELVING 36 50 80 1380 ≤ 1380 cells
ASSEMBLY 22.5 25 40 760.7143 ≤ 900
BAS/STD MIX -0.3333 1 11.04814 ≥ 0 Objective: 100 120
162.8571
3525.714 max
GROSSPROF 4
BEFORE CUTTING 16 12 20 480 ≤ 480
WELDING 25 22.5 36 757 ≤ 900
PAINTING 1 1 1 32 ≤ 32
Constraints
SHELVING 36 50 80 1380 ≤ 1380
ASSEMBLY 22.5 25 40 780 ≤ 900
BAS/STD MIX -0.3333 1 3 ≥ 0
33
Sensitivity report bagian variable cells menunjukkan seberapa besar nilai koefisien fungsi objektif yang dapat diubah
tanpa mempengaruhi nilai optimal variabel keputusan.

Contoh: untuk produk jenis Basic, gross profit per unit dapat berubah dalam rentang 100-13,6 dan 100+1,33 tanpa
mengubah nilai optimal variabel keputusan yang sekarang.

34
 Shadow Price
Sensitivity report bagian constraint menunjukkan nilai
shadow price, selisih nilai optimal fungsi objektif jika
nilai RHS pada constraint ditambahkan 1 satuan pada
constraint yang bersifat binding.

Contoh:
Penambahan kapasitas waktu overtime operasi
shelving dari 1380 menit menjadi 1381 menit
Shadow price = 3527,143 – 3525,714 = $1,428571
Decision variables
BASIC STANDARD LUXURY Sum sign RHS
Changing cells 15.64286 16.35714 0 Nilai optimal fungsi objektif akan naik
Objective:
GROSSPROF
100 120 160 3527.143 max sebesar $1,428571 pada setiap penambahan 1
CUTTING 16 12 20 446.5714 ≤ 480 menit nilai kapasitas waktu overtime operasi shelving
WELDING 25 22.5 36 759.1071 ≤ 900 sampai 220 menit dan akan turun juga
PAINTING 1 1 1 32 ≤ 32 sebesar $1,428571 pada setiap pengurangan 1
Constraints SHELVING 36 50 80 1381 ≤ 1381 menit nilai kapasitas waktu overtime operasi shelving
ASSEMBLY 22.5 25 40 760.8929 ≤ 900
hingga 116 menit
BAS/STD MIX -0.3333 1 11.14338 ≥ 0
35
Limits Report
36
 Limit Reports
Menunjukkan bagaimana nilai fungsi
objektif berubah jika setiap variabel
keputusan berada pada nilai
terendah atau tertinggi yang
mungkin terjadi (lowest or highest
feasible value).

Decision variables
Contoh: Jika variabel keputusan BASIC STANDARD LUXURY Sum sign RHS
produk jenis STANDARD ditetapkan
menjadi 5,2376 (nilai minimumnya Changing cells 15.71429 5.23760 0
yaitu, dari 1/3 variabel keputusan Objective: GROSSPROF 100 120 160 2199.940571 max
produk jenis BASIC), nilai fungsi
objektif akan menjadi $2199,94 CUTTING 16 12 20 314.2797714 ≤ 480
WELDING 25 22.5 36 510.7031429 ≤ 900
PAINTING 1 1 1 20.95188571 ≤ 32
Constraints SHELVING 36 50 80 827.5942857 ≤ 1380
ASSEMBLY 22.5 25 40 484.5114286 ≤ 900

BAS/STD MIX -0.3333 1 2.85714E-05 ≥ 0

37
Number of Binding Constraints
and Positive Optimal Values of
Decision Variables
• The number that have positive optimal values is never larger
than the number of constraints that are binding.
• Even if the choice of workstation types had been in the
hundreds, at most six would ever have been chosen as
optimal for production if only six constraints restricted their
values. And if one or several constraints had been slack, the
solution would have used even fewer types.
• The number of constraints dictates the variety of choice that
is optimal.

39
 Unsuccessful solver
• Solver may terminate without finding the solution
• No combination of decision variable values exists that satisfies all constraint -> Solver couldn’t find
a feasible solution
• Unbounded solution -> TheSetCell values do not converge (the constraint set may increased the
decision variables value without upperbound and never be restricted by a constraint -> The
objective function value can go to plus infinity)
• Computer input
• - Wrong sign (+/- and >/<) -> variable go infinity
• - Left some LHS coeff at zero
• - Wrong column/row at inputing
• - Wrong spelling the variable

40
14.5

Effect of Forcing Production of

Luxury
• The reduction cost for LUXURY only -2.85714. Forcing the production of some LUXURY units may
cause only a small decrease of daily gross profit, while giving the firm a more balanced
product line. Management may consider that not offering any Luxury workstation may be
detrimental to its market image.
• Add constraint :

• The result :

42
• The result before adding the constraint :

43
LIMITATIONS OF SENSITIVITY ANALYSIS
• No information given on the permissible increase in the decision variable for
which it is valid
• Assumption that only input parameter or coeff is changed at a time, with all
other inputs remaining at their original values
• For instance in the product-mix problem, an increase in raw material costs will
affect every single objective function coefficient. Similarly, management may fee
obliged to offer overtime to all workers and not just those making the shelves.
Again this means that several RHS parameters will change simultaneously.
• The Sensitivity Report’s main value is to highlight further aspects for analysis.

44
Pineapple
Delight
Case Study

14.6

45
Situation Summary
• Pineaple Delight Inc. memiliki pineaple plantaion
• Harga pineaple dijual $440 per ton
• Estimasi hasil musim ini 24.000 ton
• Processing
• Skinned (waste: 39,4%)
• Outer & core jadi juice (26,6%)
• Flesh (34%)
• Flesh (dulunya) jadi ring shape
• Karena kompetisi tambah juga menjadi spear, chunks shapes
• Chunks shapes dijual langsung atau digabung dengan delima
• Offcuts (potongan kecil/crush): fruit salad, fruit catering, crush
catering packs

46
Additional Information • Rings delight uses 81% Flesh
rings

• 1 kaleng rings delight seberat

454 gram,
• 70% berat bersih daging nanas • Spears delight uses 60% flesh
= 317 grams spears

Offcuts = Sisa masing-masing flesh

• Chunks delight uses 28%
• Rings : (100-81)% = 19% flesh spears dan 85% chunks
• Spears: (100-60-28) %= 12% flesh
• Chunks : (100-85) %= 15%

47
Product Information
48
 • Menyusun rencana operasi untuk
musim tersebut
• Objective : Maksimasi profit
Keinginan
• Menentukan :
Manajemen??? • Jumlah Pineapple Fruit dijual
• Alokasi penjualan daging nanas
setelah di proses

49
Creating a
Material Flow
Diagram

50
Stages Pineaple
input

Sold to Used in
other canner
firms y

Used in
cannery

Flesh Waste &

Availables Juice

51
Stages
Flesh
Availables

Flesh for Flesh for Flesh for

Rings Spears Chunks

52
Stages
Flesh for Flesh for Flesh for
Rings Spears Chunks

Chunks
Rings Spears
In Offcuts
Delight Delight
Progress

Passion
Chunks
Fruit Mix
Delight
Deligth

53
Stages
Offcuts

Salad Crush
Fruid
Catering Catering
Salad
Packs packs

54
Final Material
Flow Diagram

55
 Variable Units and Names
VARIABEL DESKRIPSI

Units sold jumlah nanas yang dijual sebagai buah

used jumlah nanas yang diolah lebih lanjut

Scaling flesh total daging nanas untuk diolah

1:1000 fleshrings daging untuk diolah menjadi ring

• 1000 cans fleshspears daging untuk diolah menjadi spears

fleshchunk daging untuk diolah menjadi chunks

• 1000 kg
rings total kaleng ring yang terjual

spears total kaleng spears yang terjual

chunkswip chunks yang dihasilkan

Why??? chunks output of pineapple chunks

pfchunks total produk chunks dan passion fruit

- Computation Accuracy offcuts sisa yang dihasilkan

occrushpacks output offcut catering crush

- Simplicity
ocsaladcater output offcut salad catering packs

ocfruitsalad output offcut fruit salad 56

57
Objective Functions

Harga pineaple dijual $440 per ton

Maksimasi

0,44*sold+0,62*rings+0,69*spears+0,69*pfchunks +
0,58*chunks + 0,41*ocfruitsalad +2,15*ocsaladcater +
4,22*occrushpacks

58
 Processing Constraints

59
 Pineapple
input
sold + used <= 24000

Sold to Used in
other canner
firms y

Estimasi hasil musim ini 24.000 ton

Used in
canner used*0,34 = flesh
y

Flesh Flesh (34% Pineaple Used)

Available
s
60
 Flesh
Available
s

Flesh for Flesh for Flesh for

Rings Spears Chunks

flesh = fleshrings+fleshspears+fleshchunk
Flesh for
Rings
Rings delight uses 81% Flesh rings

Rings
Delight
fleshrings*0,81 = 0,31*rings 61
fleshspeares*0,60 = 0,27*spears
fleshchunks*0,85+ 28%fleshspears = 0,85*chunkswip

Chunks
In
Progress

Passion
Chunks
Fruit Mix
Delight
Deligth

chunkswip = 0.34*chunks + 0.31*pfchunks

62
Rings : (100-81)% = 19%
Spears: (100-60-28) %= 12%
Chunks : (100-85) %= 15%

0,19*fleshrings+0,15*fleshchunks+ 0,12*fleshspears = offcuts

63
 Offcuts

Salad Crush
Fruid
Catering Catering
Salad
Packs packs

offcuts = 3.4*occrushpack + 1,6*ocsaladcater + 0,16*ocfruitsalad

64
 Processing Cosntraint

sold + used <= 24000

used*0,34 = flesh
flesh = fleshrings+fleshspears+fleshchunk
fleshrings*0,81 = 0,31*rings
fleshspeares*0,60 = 0,27*spears
fleshchunks*0,85+ 28%fleshspears = 0,85*chunkswip
19%*fleshrings+15%fleshchunks+ 12%fleshspears = offcuts
chunkswip = 34%*chunks + 31%*pfchunks
3.4*occrushpack + 1,6*ocsaladcater +
offcuts = 0,16*ocfruitsalad

65
Marketing
Constraint
rings <= 9000
spears <= 6000
pfchunks <= 4000
chunks <= 8000
ocfruitsalad <= 7000
ocsaladcater <= 3000

66
 Bentuk penyelesaian pada excel
definisi variabel keputusan man made constraint
sold 0
used 0 number of rings allowed 0 <= 9000
flesh 0
fleshrings 0
fleshspears 0 number of spears allowed 0 <= 6000
fleshchunk 0
rings 0
spears 0 numbers of passion fruit allowed 0 <= 4000
chunkswip 0
chunks 0
pfchunks 0
offcuts 0 number of chunks allowed 0 <= 8000
occrushpacks 0
ocsaladcater 0
ocfruitsalad 0 numbers of fruit salad allowed 0 <= 7000

numbers of fruit salad cater

0 <= 3000
allowed

67
 system constraint
Constraint type LHS symbol RHS
Bentuk total harvest 0<= 24000

penyelesaian pada usable flesh 0= 0

excel flesh allocation 0= 0

rings produced 0= 0

objective function spears produced 0= 0

0

chunks available 0= 0

Dengan fungsi objektif : offcuts available 0= 0

0,44*sold+0,62*rings+0,69*spears+
0,69*pfchunks + 0,58*chunks + offcuts allocation 0= 0
0,41*ocfruitsalad+2,15*ocsaladcater
chunks allocation 0= 0
+4,22+occrushpacks

68
 variabel
definisi
keputusan
sold 0
used 24000
xxx
flesh 8160
fleshrings 3444,444444
fleshspears 2700
Optimal fleshchunk
rings
2015,555556
9000
Solutions spears
chunkswip
6000
2469,222222
chunks 3615,359477
pfchunks 4000
offcuts 1280,777778
occrushpacks 0
ocsaladcater 100,4861111
ocfruitsalad 7000
69
 Analisis Sensitivitas variabel keputusan
Objective
Name Final value Reduced cost coefficient Allowable Increase Allowable Decrease
sold variabel keputusan 0 -0,12153125 0,44 0,12153125 1E+30
used variabel keputusan 24000 0 0 1E+30 0,12153125
flesh variabel keputusan 8160 0 0 1E+30 0,357444853
fleshrings variabel keputusan 3444,444444 0 0 1E+30 0,22375
fleshspears variabel keputusan 2700 0 0 1E+30 0,520667892
fleshchunk variabel keputusan 2015,555556 0 0 0,22375 0,357444853
rings variabel keputusan 9000 0 0,62 1E+30 0,085632716
spears variabel keputusan 6000 0 0,69 1E+30 0,234300551
chunkswip variabel keputusan 2469,222222 0 0 0,263235294 0,420523356
chunks variabel keputusan 3615,359477 0 0,58 0,0895 0,142977941
pfchunks variabel keputusan 4000 0 0,69 1E+30 0,161176471
offcuts variabel keputusan 1280,777778 0 0 17,35559641 2,382965686
occrushpacks variabel keputusan 0 -0,34875 4,22 0,34875 1E+30
ocsaladcater variabel keputusan 100,4861111 0 2,15 1,95 0,164117647
ocfruitsalad variabel keputusan 7000 0 0,41 1E+30 0,195

70
 Analisis sensitivitas batasan
Objective
Name Final value Reduced cost coefficient Allowable Increase Allowable Decrease
number of rings allowed 9000 0,085632716 9000 3778,633776 4582,656546
number of spears allowed 6000 0,234300551 6000 4479,012346 5811,999134
numbers of passion fruit allowed 4000 0,161176471 4000 3965,232975 4000
number of chunks allowed 3615,359477 0 8000 1E+30 4384,640523
numbers of fruit salad allowed 7000 0,195 7000 1004,861111 7000
numbers of fruit salad cater allowed 100,4861111 0 3000 1E+30 2899,513889
total harvest 24000 0,56153125 24000 5158,400615 3152,505447
usable flesh 0 -1,6515625 0 1071,851852 1753,856209
flesh allocation 0 -1,6515625 0 1071,851852 1753,856209
rings produced 0 -1,723765432 0 1171,376471 1420,623529
spears produced 0 -1,687775735 0 1209,333333 1569,239766
chunks available 0 -1,705882353 0 1229,222222 1490,777778
offcuts available 0 -1,34375 0 160,7777778 4639,222222
offcuts allocation 0 -1,34375 0 160,7777778 4639,222222
chunks allocation 0 -1,705882353 0 1229,222222 1490,777778

71
 Analisis batasan sistem
Name Cell Value Status Slack
number of rings allowed 9000Binding 0
number of spears allowed 6000Binding 0
numbers of passion fruit allowed 4000Binding 0
number of chunks allowed 3615,359477Not Binding 4384,640523
numbers of fruit salad allowed 7000Binding 0
numbers of fruit salad cater allowed 100,4861111Not Binding 2899,513889
total harvest 24000Binding 0
usable flesh 0Binding 0
flesh allocation 0Binding 0
rings produced 0Binding 0
spears produced 0Binding 0
chunks available 0Binding 0
offcuts available 0Binding 0
offcuts allocation 0Binding 0
chunks allocation 0Binding 0

72
Keputusan yang Variabel
Nanas dijual langsung
jumlah satuan
0ton
akan diambil Nanas diolah lebih lanjut
Daging nanas
24000ton
8160ton

Nanas untuk bentuk rings 3444,44444 ton

Nanas untuk bentuk
spears 2700ton
Nanas untuk bentuk
chunks 2015,55556 ton
Kaleng bentuk rings 9000000kaleng
Profit yang didapatkan dari Kaleng bentuk spears 6000000kaleng
keputusan tersebut = Jumlah chunks yang
diproduksi 2469,22222 ton
17662953,6 $ Kaleng bentuk chunks 3615360kaleng
Kaleng passion fruit dan
chunks 4000000kaleng
Sisa nanas 1280,77778 ton
Crush catering packs 0kaleng
Fruit salad catering 100486kaleng
Fruit salad 7000000kaleng
73
Analisis lanjutan • Alokasi dari fruit salad, fruit salad
catering packs dan fruit salad crush.
• Alokasi batasan dari rings, spears, dan
chunks.
• Kemungkinan menjual buah secara
langsung.
• Kemungkinan menambah nanas.

74
 14.7

A transportation
problem
 A production/transportation problem
situation

Tomato processing plant Tomato conversion plants

● Stockton ● Los Angeles

● Riverside, California Heinz ●
●
Dallas
Chicago
● Tuscaloosa, Alabama
● Atlanta
● Newark
 Cost summarizes, capacity,
and requirements

Production capacity Transportation cost to destinations

Sources
cost (tonnes) LA Dallas Chicago Atlanta Newark
$ $ $ $ $
Stockton $524 4600
25 48 54 67 75
$ $ $ $ $
Riverside $541 2900
11 44 57 61 81
$ $ $ $ $
Tuscaloosa $612 1700
57 33 32 10 36
Amount required in tonnes 2100 1700 1600 1300 2200
 Goods flow diagram

Goods flow diagram for transportation problem. (2005). Dalam H. G. Daellenbach, & D. C. McNickle, Management
science : Decision making through systems thinking (hal. 388). New York: PALGRAVE MACMILLAN.
 Total cost per tonne to
destinations

capacity Total relevant cost to destinations (per tonnes)

sources
(tonnes) LA Dallas Chicago Atlanta Newark
$ $ $ $ $
Stockton 4600
549 572 578 591 599
$ $ $ $ $
Riverside 2900
552 585 598 602 622
$ $ $ $ $
Tuscaloosa 1700
669 645 644 622 648
Amount required in tonnes 2100 1700 1600 1300 2200
 Constraints
Supply constraint
𝑆𝑇𝐿𝐴 + 𝑆𝑇𝐷𝐴 + 𝑆𝑇𝐶𝐻 + 𝑆𝑇𝐴𝑇 + 𝑆𝑇𝑁𝐸 ≤ 4600
𝑅𝐼𝐿𝐴 + 𝑅𝐼𝐷𝐴 + 𝑅𝐼𝐶𝐻 + 𝑅𝐼𝐴𝑇 + 𝑅𝐼𝑁𝐸 ≤ 2900
𝑇𝑈𝐿𝐴 + 𝑇𝑈𝐷𝐴 + 𝑇𝑈𝐶𝐻 + 𝑇𝑈𝐴𝑇 + 𝑇𝑈𝑁𝐸 ≤ 1700

Demand constraint
𝑆𝑇𝐷𝐴 + 𝑅𝐼𝐷𝐴 + 𝑇𝑈𝐷𝐴 = 1700
𝑆𝑇𝐶𝐻 + 𝑅𝐼𝐶𝐻 + 𝑇𝑈𝐶𝐻 = 1600
𝑆𝑇𝐴𝑇 + 𝑅𝐼𝐴𝑇 + 𝑇𝑈𝐴𝑇 = 1300
𝑆𝑇𝑁𝐸 + 𝑅𝐼𝑁𝐸 + 𝑇𝑈𝑁𝐸 = 2200

Non-negativity constraints
𝑆𝑇𝐿𝐴, 𝑆𝑇𝐷𝐴, 𝑆𝑇𝐶𝐻, 𝑆𝑇𝐴𝑇, 𝑆𝑇𝑁𝐸, 𝑅𝐼𝐿𝐴, 𝑅𝐼𝐷𝐴, 𝑅𝐼𝐶𝐻,
𝑅𝐼𝐴𝑇, 𝑅𝐼𝑁𝐸, 𝑇𝑈𝐿𝐴, 𝑇𝑈𝐷𝐴, 𝑇𝑈𝐶𝐻 , 𝑇𝑈𝐴𝑇, 𝑇𝑈𝑁𝐸 ≥ 0
 Objective function

𝑀𝑖𝑛 𝑍 = 549 𝑆𝑇𝐿𝐴 + 572 𝑆𝑇𝐷𝐴 + 578 𝑆𝑇𝐶𝐻

+ 591 𝑆𝑇𝐴𝑇 + 599 𝑆𝑇𝑁𝐸 + 552 𝑅𝐼𝐿𝐴
+ 585 𝑅𝐼𝐷𝐴 + 598 𝑅𝐼𝐶𝐻 + 602 𝑅𝐼𝐴𝑇
+ 622 𝑅𝐼𝑁𝐸 + 669 𝑇𝑈𝐿𝐴 + 645 𝑇𝑈𝐷𝐴
+ 644 𝑇𝑈𝐶𝐻 + 622 𝑇𝑈𝐴𝑇 + 648 𝑇𝑈𝑁𝐸
 Optimal Solution

Destination Transportation schedule (tonnes) Total Capacity

Sources LA Dallas Chicago Atlanta Newark shipped (tonnes)
Stockton 0 900 1600 0 2100 4600 4600
Riverside 2100 800 0 0 0 2900 2900
Tuscaloosa 0 0 0 1300 100 1400 1700
Total received 2100 1700 1600 1300 2200
Ammount Total Cost $ 5,198,100
2100 1700 1600 1300 2200
Required

$5,198,100
Other problem situations disguised
as transportation problems

regular time/overtime assignment

production scheduling problem.
problem.
Terima Kasih