Operations Management

P6 - Queueing Systems

João Ramadas
[email protected]
 Practical Class #6 - Summary

Any questions from last class?

Topic 3 – Queueing Systems

• Exercise 3.1
• Exercise 3.3
• Exercise 3.4

Operations Management – João Ramadas 2

Operations Management – João Ramadas 3
 Queueing systems | Assumptions

Queueing Assumptions
• FIFO (First in First out)
• No bulking or reneging (everyone that sees a line is patient and stays in line)
• Arrivals are independent
• Service times are independent
• Service rates and arrival rates are steady

Operations Management – João Ramadas 4

 Queueing systems | Ex. 3.1

Quick Lube Inc. operates a fast lube and oil change garage. On a typical day, customers arrive at the rate of 3 per
hour, and lube jobs are performed at an average rate of one every 15 minutes. The mechanics operate as a team on
one car at a time.

Assumptions: Poisson arrivals and exponential service.

1. What is the model?

Single server – single phase

Queue Server Departure

Arrival

Operations Management – João Ramadas 5

 Queueing systems | Ex. 3.1

Quick Lube Inc. operates a fast lube and oil changλe garage. On a typical day, customers arrive at the rate of 3 per
hour, and lube jobs are performed at an average rate of one every 15 minutes. The mechanics operate as a team on
one car at a time.

Assumptions: Poisson arrivals and exponential service.

2. What is the Kendal's notation?

The arrival rate: Poisson distribution (M) λ= 3 customers per hour

The service rate: Exponential distribution (M) 𝜇= 4 customers per hour
The number of servers: One (1)

Operations Management – João Ramadas 6

 Queueing systems | Ex. 3.1

Quick Lube Inc. operates a fast lube and oil changλe garage. On a typical day, customers arrive at the rate of 3 per
hour, and lube jobs are performed at an average rate of one every 15 minutes. The mechanics operate as a team on
one car at a time.

Assumptions: Poisson arrivals and exponential service.

Before moving forward…  

Operations Management – João Ramadas 7

 Formulas for MM1

Avg. server utilisation

Avg. # of customers in Queue

Avg. # of customers in System

Avg. time a customer spends waiting in Queue

Avg. time a customer spends waiting in System

Probability of zero customers

Probability of n customers

Operations Management – João Ramadas 8

 Queueing systems | Ex. 3.1 a) and b)

a. What is the utilization of the lube team?

Model: MM1

  λ = 3 customers per hour

𝜇 = 4 customers per hour

Answer:

b. What is the average numbers of cars in line?

 
Answer:

Operations Management – João Ramadas 9

 Queueing systems | Ex. 3.1 c) and d)

c. What is the average time a car waits before it is lubed?

Model: MM1

  Answer:

d. What is the total time it takes to go through the system? (that is, waiting in line plus lube time)

  Answer:

 
or…  

Operations Management – João Ramadas 10

 Queueing systems | Ex. 3.3

The Bijou Theatre in Hermosa Beach, California, shows vintage movies. Customers arrive at the theater line at the
rate of 100 per hour. The ticket seller averages 30 seconds per customer. This includes placing validation stamps on
customers’ parking lot receipts and punching their frequent watcher cards.

Assumptions: Poisson arrivals and exponential service.

1. What is the model? Single server – single phase

2. What is the Kendal's notation? MM1

3. What is the arrival rate (λ): 100 customers per hour

4. What is the service rate (𝜇): 60/0,5 = 120 customers per hour

Operations Management – João Ramadas 11

 Queueing systems | Ex. 3.3 a) and b)

a. What is the average time a customer waits before he sits?

Model: MM1

λ = 100 customers per hour

 
𝜇 = 60/0,5 = 120 customers per hour

Answer:

b. What would be the effect of having a second ticket seller doing nothing but the validations and card punching,
thereby cutting the average service time to 20 seconds?
Does this change our model? What changes?

λ = 100 customers per hour

 
𝜇 = 60/0,3(3) = 180 customers per hour

Answer:

Operations Management – João Ramadas 12

 Queueing systems | Ex. 3.3 c)

c. Would W be less than in b) if a 2nd window was opened with each server doing all the tasks?
Does this change our model? What changes?
Assuming 1 single queue

1. What is the model? Multiple servers – single phase

2. What is the Kendal's notation? MM2

3. What is the arrival rate (λ): 100 customers per hour

4. What is the service rate (𝜇): 60/0,5 = 120 customers per hour
New situation:

- Lq is now very complicated to calculate… so we use ρ and the Lq table

Operations Management – João Ramadas 13

 Formulas for MMC

Avg. server utilisation

Avg. # of customers in Queue

Avg. # of customers in System

Avg. time a customer spends waiting in Queue

Avg. time a customer spends waiting in System

Probability of zero customers

Probability of n customers

Operations Management – João Ramadas 14

 Queueing systems | Lq table

Operations Management – João Ramadas 15

 Queueing systems | Ex. 3.3 c)

c. Would W be less than in b) if a 2nd window was opened with each server doing all the tasks?
Assuming 1 single queue

λ = 100 customers per hour

𝜇 = 60/0,5 = 120 customers per hour
C = 2 servers

     

Operations Management – João Ramadas 16

 Queueing systems | Lq table

Operations Management – João Ramadas 17

 Queueing systems | Ex. 3.3 c)

c. Would W be less than in b) if a 2nd window was opened with each server doing all the tasks?
Assuming 1 single queue

λ = 100 customers per hour

𝜇 = 60/0,5 = 120 customers per hour
C = 2 servers

     

   
 

Operations Management – João Ramadas 18

 Queueing systems | Ex. 3.3 c)

What if we consider 2 queues?

1. What is the model? Single server – single phase

2. What is the Kendal's notation? 2 x MM1

3. What is the arrival rate (λ): 100/2 = 50 customers per hour

4. What is the service rate (𝜇): 60/0,5 = 120 customers per hour

   

Operations Management – João Ramadas 19

 Queueing systems | Ex. 3.4

In order to improve customer service at ZMarket, its management wants to test two alternative options to reduce
customer waiting time: recruiting one extra employee to pack up purchases; or add one checkout counter.
ZMarket has only one checkout counter. Customers arrive at ZMarket at a rate of 24 customers/hour and the current
service rate at the checker is 30 customers per hour. It is known that adding one extra employee would increase
service rate from 30 to 40 customers per hour.

Consider the following:

- The cost of one extra employee is $150/week.
- Each one-minute reduction in customer waiting time avoids $75 / per week in lost sales.
- One new counter costs $6.000 plus $200 per week for checker.
- Customers divide themselves between two checkout lines.
- The service rate (μ) for each checker is 30

a) Analyze the problem and indicate what the best alternative is.

Operations Management – João Ramadas 20

 Queueing systems | Ex. 3.4

Currently +1 employee ($150/week) +1 checkout ($200/week)

Kendal's notation? Kendal's notation? Kendal's notation?

MM1 MM1 2 x MM1

Arrival Rate? Arrival Rate? Arrival Rate?

λ = 24 customers/hr λ = 24 customers/hr λ = 24/2 = 12 customers/hr

Service Rate? Service Rate? Service Rate?

𝜇 = 30 customers/hr 𝜇 = 40 customers/hr 𝜇 = 30 customers/hr

Operations Management – João Ramadas 21

 Queueing systems | Ex. 3.4

MM1 MM1 2xMM1

Currently +1 employee ($150/week) +1 checkout ($200/week)

λ = 24 customers per hr λ = 24 customers per hr λ = 12 customers per hr
𝜇 = 30 customers per hr 𝜇 = 40 customers per hr 𝜇 = 30 customers per hr

   

  Comparing benefits (per week):

Costs: $200
Comparing benefits (per week): Savings: (8 - 1,3(3)) × $75 = $500

Costs: $150
Added profits = $300
Savings: (8 - 2,25 min) × $75 = $431,25

Added profits = $281,25

Can we make an advice based on these profits?

Operations Management – João Ramadas 22

 Queueing systems | Ex. 3.4

MM1 MM1 2xMM1

Currently +1 employee ($150/week) +1 checkout ($200/week)

   

Comparing benefits (per week): Comparing benefits (per week):

Costs: $150 Costs: $200
Savings: (8 - 2,25 min) × $75 = $431,25 Savings: (8 - 1,3(3)) × $75 = $500

Added profits = $281,25 Added profits = $300

Considering investment in counter ($6.000)

Operations Management – João Ramadas 23

 Queueing systems | Ex. 3.4

Conclusions

- Hiring a new worker increase profits and increase service quality

- Adding a new counter increases profit a little bit more and increases also service quality, but only after initial
investment being covered

- One extra counter can give more idle time to employees

- One extra counter can give the business less flexibility

- One extra counter will occupy precious space in the store

Operations Management – João Ramadas 24

 End of Practical Class 6…

Thank you!

Operations Management – João Ramadas 25