Part 3 Simulation with R

Dr. Nguyen Quang Huy

Dec 30, 2019

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 1 / 35
Why simulation.

Real world problem

Problem designed SIMULATION

Computer calculation

Real world solution

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 2 / 35
Why simulation

Real world problem

(uncertainty output)

RANDOM
SIMULATION

Real world solution

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 3 / 35
Monte Carlo simulation
Monte Carlo (MC) simulation is based on the law of large number in
probability theory:
Law of large number Let X1 , X2 , · · ·, is an infinite sequence of i.i.d
random variables with expected value E (X1 ) = E (X2 ) = · · · = µ; the
sample average

1
X̄n = (X1 + X2 + · · · + Xn )
n

converges to the expected value

X̄n → µ for n → ∞

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 4 / 35
Why simulation
Example 1: Flip 100 fair coins simutaneously, how many coins show head
(H) ? → design problem on computer:

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 5 / 35
Why simulation
Example 1: Flip 100 fair coins simutaneously, how many coins show head
(H) ? → design problem on computer:

x<-sample(c(0,1),100,replace=TRUE) #goolge for sample function

sum(x) # real world answer

## [1] 49

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 5 / 35
Why simulation
Example 1: Flip 100 fair coins simutaneously, how many coins show head
(H) ? → design problem on computer:

x<-sample(c(0,1),100,replace=TRUE) #goolge for sample function

sum(x) # real world answer

## [1] 49
Example 2: Flip 100 fair coins simutaneously, what the probability that the
numbers of coin showing head (H) is larger than 65

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 5 / 35
Why simulation
Example 1: Flip 100 fair coins simutaneously, how many coins show head
(H) ? → design problem on computer:

x<-sample(c(0,1),100,replace=TRUE) #goolge for sample function

sum(x) # real world answer

## [1] 49
Example 2: Flip 100 fair coins simutaneously, what the probability that the
numbers of coin showing head (H) is larger than 65

y<-vector(mode="numeric",10^6)
for (i in 1:10^6){x<-sample(c(0,1),100,replace=TRUE)
y[i]<-sum(x)}
sum(y>65)/10^6

## [1] 0.00087
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 5 / 35
Why simulation

Example 3: Rolling 2 dies simutaneously, what is probability that the sum of

2 faces is less than or equal to 5 → EASY . . .

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 6 / 35
Why simulation

Example 3: Rolling 2 dies simutaneously, what is probability that the sum of

2 faces is less than or equal to 5 → EASY . . .

11 12 13 14 15 16
21 22 23 24 25 26
31 32 33 34 35 36
41 42 43 44 45 46
51 52 53 54 55 56
61 62 63 64 65 66

10 5
→ answer is =
36 18

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 6 / 35
Why simulation
Example 3 (continue): Rolling 10 dies simutaneously, what is probability
that the sum of 10 faces is less than or equal to 30 → NOT EASY →
SIMULATION

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 7 / 35
Why simulation
Example 3 (continue): Rolling 10 dies simutaneously, what is probability
that the sum of 10 faces is less than or equal to 30 → NOT EASY →
SIMULATION

y<-vector(mode="numeric",10^6) # rolling 10 dies 1 million tim

for (i in 1:10^6){x<-sample(c(1,2,3,4,5,6),10,replace=TRUE)
y[i]<-sum(x)}
sum(y<=30)/10^6

## [1] 0.204038
Example 3 (continue): Rolling 10 dies simutaneously. Finding n such that

n = arg max P (sum of 10 faces = k)

6≤k≤60

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 7 / 35
Why simulation0.06
0.04
Probability

0.02
0.00

0 10 20 30 40 50 60

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 8 / 35
Why simulation
Example 4 In a card game where 52 cards are dealt evenly to 4 players.
What is probability of Quads (Tu quy)

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 9 / 35
Why simulation
Example 4 In a card game where 52 cards are dealt evenly to 4 players.
What is probability of Quads (Tu quy)

x<-c(1:13,1:13,1:13,1:13)
TestTuQuy<-function(y){
h<-vector(mode="numeric",13)
for (i in 1:13){h[i]<-sum(y==i)}
ifelse(max(h)==4,1,0) }
Total<-0
for (i in 1:10){#chia bai 10 lan
j<-sample(1:52)
z<-x[j]
if (TestTuQuy(z[1:13])+TestTuQuy(z[14:26])+TestTuQuy(z[27:39])
Total

## [1] 2
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 9 / 35
Simulation to solve probability problem
Problem 1: Blind spider
An ant and a blind spider are on opposite corners of a cube. The ant is
stationary and the spider moves at random from one corner to another
along the edges only. What is the expected number of turns before the
spider reaches the ant ?

Một con kiến và một con

nhện mù ở hai góc đối diện
của một hình lập phương.
Con kiến là
đứng yên và con nhện di
chuyển ngẫu nhiên từ góc
này sang góc khác
chỉ dọc theo các cạnh. Số
lượt dự kiến ​trước khi
nhện đến được kiến?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 10 / 35
Simulation to solve probability problem

Problem 1: Blind spider Vị trí hiện tại của con nhện là (0,0,0) trong khi vị trí của con kiến là
(1,1,1).
Current position of the spider is (0,0,0) while position of the ant is
(1,1,1).
How does the spider randomly move ?
Current position of the spider is (x,y,z) with x , y , z ∈ {0, 1}
Next position of the spider is determined by randomly choosing between
x, y, z and replace it by (1 - value of itself).
For example, the current position of the spider is (0,1,0). The sample
function sample(1:3) returns value 1, the spider is moving as follows:
trả về giá trị 1
(0, 1, 0) → (1, 1, 0)

The spider reaches the ant if the current position of the spider is (1,1,1)
Con nhện đến chỗ con kiến nếu vị trí hiện tại của con nhện là (1,1,1)

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 11 / 35
Simulation to solve probability problem

Problem 1: Blind spider How many steps that the spider needs to reach
the ant:

SpiderPosition<-c(0,0,0)
AntPosition<-c(1,1,1)
MovingNumber<-0
while(sum(abs(SpiderPosition-AntPosition))>0){
k<-sample(1:3,1)
SpiderPosition[k]<-1-SpiderPosition[k]
# print(SpiderPosition) # if you want to see the spider po
MovingNumber<-MovingNumber+1
}
MovingNumber

## [1] 3
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 12 / 35
Simulation to solve probability problem
Problem 1: Blind spider Let the spider catches the ant 100,000 times (or
more) to calculate the expected number of movement

Result<-vector(mode="numeric",10^5)
for (i in 1:10^5){
SpiderPosition<-c(0,0,0)
AntPosition<-c(1,1,1)
MovingNumber<-0
while(sum(SpiderPosition==AntPosition)<3){
k<-sample(1:3,1)
SpiderPosition[k]<-1-SpiderPosition[k]
MovingNumber<-MovingNumber+1}
Result[i]<-MovingNumber}
mean(Result)

## [1] 9.96394
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 13 / 35
Simulation to solve probability problem
Problem 1: Blind spider (continue) We can calculate the probability
mass function of number of movement:
0.20
0.15
Probability

0.10
0.05

5 10 15 20

Number of movement
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 14 / 35
Simulation to solve probability problem
Problem 2 - Monkey typing: Suppose a monkey is typing randomly at a
typewriter whose only keys are the capital letters A, D, M, U. What is the
expected time it will take for the monkey to type the following words

"MAD" giả sử một con khỉ đang gõ ngẫu nhiên vào một
"MUM" máy đánh chữ có phím duy nhất là các chữ in hoa A, D, M, U.
Cái gì là
thời gian dự kiến để con khỉ gõ những dòng sau

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 15 / 35
Simulation to solve probability problem

KeyBoard<-c("A","D","M","U")
str<-KeyBoard[sample(1:4,3,replace=TRUE)]
MyString<-paste(str[1],str[2],str[3])
TypingNumber<-3
while (sum(str==c("M","A","D"))<3){ str[1]<-str[2]
str[2]<-str[3]
str[3]<-KeyBoard[sample(1:4,1)]
MyString<-paste(MyString,str[3])
TypingNumber<-TypingNumber+1}
TypingNumber

## [1] 222

substr(MyString,max(0,nchar(MyString)-20),nchar(MyString))

## [1] "D A M A U D M M M A D"

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 16 / 35
Simulation to solve probability problem
Problem 2 The monkey types until the word “MUM” appears:

result<-vector(mode="numeric",10)
for (i in 1:10){#10^4 when running
KeyBoard<-c("A","D","M","U")
str<-KeyBoard[sample(1:4,3,replace=TRUE)]
MyString<-paste(str[1],str[2],str[3])
TypingNumber<-3
while (sum(str==c("M","U","M"))<3){ str[1]<-str[2]
str[2]<-str[3]
str[3]<-KeyBoard[sample(1:4,1)]
MyString<-paste(MyString,str[3])
TypingNumber<-TypingNumber+1}
result[i]<-TypingNumber}
mean(result) # "MUM": 68 and "MAD": 64

## [1] 92.5
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 17 / 35
Simulation to solve probability problem
Problem 3 - Drunk passenger(s): You are at the end of a line of 100
airline passengers is waiting to board a plane. The nth passenger in line has
ticket for the seat number n. Being drunk, the first person in line picks a
random seat (equally likely for each seat). All of the other passengers will
go to their proper seats unless it is already occupied; If it is occupied, they
will then find a free seat to sit in, at random. What is the probability that
you will sit in your proper seat ? What does this probability change if there
are k drunk passengers ?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 18 / 35
Simulation to solve probability problem

Problem 3 - Drunk passenger(s)

Step 1: First k passengers take k seats randomly.

Step 2: Let passengers (k + 1), (k + 2), · · · , 99 board respectively, if
his proper seat is occupied he will choose an empty seat randomly.
Step 3: If passenger 100 seats on 100th seat, return 1 and return 0
otherwise.
Step 4: Repeat the procedure N times (N = 105 for example) and
calculated the expectation of returns.

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 19 / 35
Simulation to solve probability problem

Problem 3 - Drunk passenger(s)

Seat<-vector(mode="numeric",100)
SeatAvailable<-1:100
Seat[1]<-sample(1:100,1) #Seat of Drunk man
SeatAvailable<-SeatAvailable[-Seat[1]] #Remove seat 1
for (i in 2:100){
if (sum(SeatAvailable==i)==0){
Seat[i]<-sample(SeatAvailable,1)}
else{Seat[i]<-i}
SeatAvailable<-SeatAvailable[-Seat[i]]# Remove seat i
}
Seat[100]

## [1] 13
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 20 / 35
Simulation to solve probability problem - Drunk
passenger
result<-vector(mode="numeric",10)#using 10^5 simulation in you
for (j in 1:10){
Seat<-vector(mode="numeric",100)
SeatAvailable<-1:100
Seat[1]<-sample(1:100,1) #Seat of Drunk man
SeatAvailable<-SeatAvailable[-Seat[1]] #Remove seat 1
for (i in 2:100){
if (sum(SeatAvailable==i)==0){
Seat[i]<-sample(SeatAvailable,1)}
else{Seat[i]<-i}
SeatAvailable<-SeatAvailable[-Seat[i]]}
result[j]<-Seat[100]}
result
## [1] 100 10 100 26 100 100 42 41 33 100
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 21 / 35
Simulation to solve probability problem

Problem 3 - Drunk passenger(s) What is the probability that you will sit
on your proper seat if 1st passenger, 2nd passenger, · · ·, k th passenger are
drunk ?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 22 / 35
Simulation to solve probability problem
Problem 4 - Messing with envelops: There are n letters and n
envelopes. Your servant puts the letters randomly in the envelopes so that
each letter is in one envelope and all envelopes have exactly one letter.
Calculate the expected number of envelopes with correct letter inside them.

Gây rối với phong bì: có n chữ cái và n

bao lì xì. Người hầu của bạn đặt các chữ cái ngẫu nhiên trong các
phong bì để
Mỗi chữ cái nằm trong một phong bì và tất cả các phong bì có chính
xác một chữ cái.
Tính số lượng phong bì dự kiến với chữ cái chính xác bên trong
chúng.

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 23 / 35
Simulation to solve probability problem
Problem 4 - Messing with envelops: There are n letters and n
envelopes. Your servant puts the letters randomly in the envelopes so that
each letter is in one envelope and all envelopes have exactly one letter.
Calculate the expected number of envelopes with correct letter inside them.

n<-20
x<-1:n
for (i in 1:10){#Using 10^6 simulations
y<-sample(1:n,n,replace=FALSE)
result[i]<-sum(x==y)}
result

## [1] 0 1 1 1 0 1 0 0 0 0

mean(result)

## [1] 0.4
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 23 / 35
 Simulation to solve probability problem
Problem 5 - Tennis tournement A tennis tournament has k = 4 levels
and n = 2k players. At the begining, random pairs are formed and one
player from each pair proceeds to next level. Given that player number i has
better skill than player j if i < j and the one with better skills always wins.
Kỹ năng tốt hơn người chơi j nếu tôi <j và người có kỹ năng tốt hơn luôn chiến thắng.
What is the probability that a player outside the top 5 plays the final
match.
Your skill is 5, calculate the probability that you will play in Semi-final.
Giải đấu quần vợt Một Xác suất người
giải đấu quần vợt có K chơi ngoài top 5
= 4 cấp độ chơi trận chung
và n = người chơi 2K. kết là gì
Khi bắt đầu, các cặp cuộc thi đấu.
ngẫu nhiên được hình Kỹ năng của bạn
thành và một là 5, tính xác suất
Người chơi từ mỗi cặp mà bạn sẽ chơi
tiến hành đến cấp độ trong trận bán
tiếp theo. Cho rằng số kết.
người chơi tôi có
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 24 / 35
Simulation to solve “tennis tournement” problem

k<-4
position<-1:(2^k)
player<-sample(1:(2^k),2^k,replace=FALSE)
for (i in 1:(k-1)){# (k-1) levels, not final match yet
OutTour<-c(0)
for (j in 1:2^(k-i)){
a<-ifelse(player[2*j-1]>player[2*j],2*j-1,2*j)
OutTour<-c(OutTour,a)}
player<-player[-OutTour[2:(2^(k-i)+1)]]
}
player # Final match

## [1] 2 1

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 25 / 35
Simulation to solve “tennis tournement” problem
k<-4
position<-1:(2^k)
result<-vector(mode="numeric",10)
for (n in 1:10){
player<-sample(1:(2^k),2^k,replace=FALSE)
for (i in 1:(k-1)){# (k-1) levels, not final match yet
OutTour<-c()
for (j in 1:2^(k-i)){
a<-ifelse(player[2*j-1]>player[2*j],2*j-1,2*j)
OutTour<-c(OutTour,a)}
player<-player[-OutTour]
}
if (max(player)>5){result[n]<-1}}
mean(result)

## [1] 0
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 26 / 35
Simulation to solve “tennis tournement” problem

Calculate the probability that you, with skill is equal to 5, will play in
Semi-final ?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 27 / 35
Simulation to solve probability problem

Problem 6 - Collecting lucky coupons: Pepsico company is holding a

contest where everyone who collects one each of 10 different coupons wins
a football ticket. You get a coupon with each purchase of a Lay’s chips
packet, and each coupon is equally likely. Whats the expected number of
packets you have to eat in order to get a ticket ?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 28 / 35
Simulation to solve probability problem
k<-5
YourCoupon<-vector(mode="numeric",5)
Eating<-c() Thu thập phiếu giảm giá may mắn:
EatNumber<-0 Công ty PepsiCo đang tổ chức
Cuộc thi mà tất cả những người thu
while(sum(YourCoupon)<5){ thập một trong số 10 phiếu giảm giá
NewOne<-sample(1:k,1) khác nhau
Eating<-c(Eating,NewOne) Một vé bóng đá. Bạn nhận được một
phiếu giảm giá với mỗi lần mua một
YourCoupon[NewOne]<-1 con chip Lay Lay
} Gói, và mỗi phiếu giảm giá có khả
length(Eating) năng như nhau. Số lượng dự kiến của
Gói bạn phải ăn để lấy vé?

## [1] 21

Eating

## [1] 5 3 3 5 3 3 1 1 5 1 3 1 5 3 1 3 3 4 5 3 2
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 29 / 35
Simulation to solve probability problem

k<-5
result<-vector(mode="numeric",10)
for (i in 1:10){#using 10^6 simulation
YourCoupon<-vector(mode="numeric",5)
Eating<-c()
EatNumber<-0
while(sum(YourCoupon)<5){
NewOne<-sample(1:k,1)
Eating<-c(Eating,NewOne)
YourCoupon[NewOne]<-1
}
result[i]<-length(Eating)}
mean(result)

## [1] 13.1
Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 30 / 35
Simulation to solve probability problem
Problem 7 - Winning a free movies ticket: At a movie theater, the
manager announces that they will give a free ticket to the first person in
line whose birthday is the same as someone who has already bought a
ticket. You have the option of getting in line at any time. Assuming that
you don’t know anyone else’s birthday. What position in line gives you the
greatest chance to win the free ticket?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 31 / 35
Simulation to solve probability problem

result<-vector(mode="numeric",10)# recommend 10^6 simulations

for(i in 1:10){
birthday<-c() tại một rạp chiếu phim,
newbirthday<-sample(1:365,1) người quản lý thông báo rằng họ
sẽ đưa vé miễn phí cho người đầu
position<-1 tiên ở hàng mà có sinh nhật giống
while(sum(birthday==newbirthday)==0){ như một người đã mua một
birthday<-c(birthday,newbirthday) vé. Bạn có tùy chọn để xếp hàng
bất cứ lúc nào. Giả sử rằng
newbirthday<-sample(1:365,1) bạn không biết sinh nhật của bất
position<-position+1} cứ ai khác . Vị trí nào trong hàng
cung cấp cho bạn
result[i]<-position} Cơ hội lớn nhất để giành được vé
result[1:5] miễn phí?

## [1] 39 14 18 17 11

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 32 / 35
Simulation to solve probability problem
Problem 8 - All die or one survive: In a room stand n = 100 armed and
angry people. At each chime of a clock, everyone simultaneously spins
around and shoots a random other person. The persons shot fall dead and
the survivors spin and shoot again at the next chime. Eventually, either
everyone is dead or there is a single survivor. What is the probability that
there will be a survivor.

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 33 / 35
Simulation to solve probability problem
Problem 9 - Inlands and bridge: There are 10 inlands I1 , I2 , · · ·, I10 , and
you want to travel from inland I1 to inland I10 . For each j = 1, 2, . . . ,9,
inlands Ij and Ij+1 are connected by 2 bridges. These bridges are looked like
to each other but if you use the bad one, it will break down immediately
and you have to swim back to inland I1 . What is expected number of
bridges you have to pass to go to inland I10 ?

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 34 / 35
Simulation to solve probability problem

Problem 10 - Points on circle: What is the probability that n random

points on a circle are lying on a half of the circle.

Dr. Nguyen Quang Huy Part 3 Simulation with R Dec 30, 2019 35 / 35