Scribd
Upload
11 views

assignment

Uploaded by

manasmi2537
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
11 views

assignment

Uploaded by

manasmi2537
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

1 11/17/23

Assignment Problem

› This is a special type of transportation problem


with sources as assignees and destinations as
tasks
› Assigning people or machine to tasks
› Basic assumptions of assignment problem:
› Number of assignees and number of tasks are the
same
› Each assignee will be given exactly one task
› Each task to be performed by exactly one
assignee
› Cost is given as cij to perform a task ‘j’ by assignee
‘i’
2 11/17/23

Assignment Problem

A1 T1

A2 T2

A3 T3

A4 T4
3 11/17/23

Assignment Problem: Formulation

› An IT firm is facing a manpower allocation


problem for two new projects: P1 and P2.
An IT professional is defined by her
knowledge (Software/Hardware) and by
her level (Developer/Manager). The
project requirements are as follows:

Knowledge Level
HW SW Developer Manager
P1 2 3 P1 2 2
P2 3 4 P2 2 3
Queueing Theory
and Applications:
Sumanta Basu
Professor
Indian Institute of Management
Calcutta
5 11/17/23

Queuing Theory: Applications

› Commercial service systems, e.g. queues


in bank counter, payment counters in mall
› Transportation service systems, e.g. cars
waiting at a tollbooth, airplanes waiting to
land or take off, parking lot etc.
› Internal service system to identify
bottleneck process, e.g. maintenance
system, inspection activity etc.
6 11/17/23

Queuing Theory: Applications

› Social service system, e.g. judicial service


systems
› Workforce planning in service sectors, e.g.
maintenance and support department in
xerox, dell etc.
› Resource planning in service operations, e.g.
restaurant, healthcare etc.
› Manpower and resource planning for
administrative, legal and emergency
activities
7 11/17/23

Queuing Theory: Decisions Taken


› Number of servers at a service facility
› Amount of waiting space in the queue
› Queuing discipline used
› Priorities for different customer categories

› Server utilization and customer waiting cost


Expected cost per unit time

Total Cost
Service Cost

Waiting Cost

Number of server(s)
8 11/17/23

Queuing Theory: Introduction

Input Service Output


Queue
Source Mechanism Source
9 11/17/23

Queuing Theory: Introduction

› Input Source (defining parameters):


› Source of input arrivals is population
› Size of the population: finite or infinite
› Distribution of arrival rate or inter-arrival time
distribution
› Any specific behavior of input source, e.g.
balking, where customers refuse to enter the
system if the queue is too long
10 11/17/23

Queuing Theory: Introduction

› Queue (defining parameters):


› Queue is where customers wait before being served
› Length of queue: finite or infinite
› Queue discipline: FIFO, LIFO or some priority
procedure

› Service Mechanism (defining parameters):


› Single or multiple servers
› Connection between servers: serial or parallel
› Distribution of service time
11 11/17/23

Notations in Queuing Theory

› λ = mean arrival rate of new customers


› µ = mean service rate
› s = number of parallel servers
› Pn = probability of exactly ‘n’ customers in
queuing system
12 11/17/23

Performance Measures in Queuing


Theory

› ρ = utilization factor = λ/(sµ)


› L = expected number of customers in
queuing system = ∑$ !"# 𝑛𝑃!
› Lq = expected number of customers in
queue = ∑$ !"%(𝑛 − 𝑠)𝑃!
› W = waiting time in the system
› Wq = waiting time in the queue
13 11/17/23

Little’s Law

› L = λW
› Lq = λWq
› W = Wq + 1/µ
14 11/17/23

Queuing Theory: Introduction

› Two random elements in queuing theory:


› Inter arrival time distribution (Exponential
distribution is the most commonly used one)
› Service time distribution (Again exponential
distribution is commonly chosen)
› Selection of distribution is a trade-off
between practical relevance and
analytical tractability
15 11/17/23

Types of Queuing Model


Service Rate (Poisson
distribution)

Number of servers: 1

Arrival Rate (Poisson


distribution)

M/M/1 System
16

M/M/1 Model

› P0 = (1-ρ) › W = 1 / (µ – λ)
› Pn = (1-ρ)ρn › Wq = λ / [µ(µ - λ)]
› L = ρ / (1-ρ)
› Lq = λ2 / [µ(µ - λ)]
17 11/17/23

Example (M/M/1)

In a retail outlet, there are three billing


counters. The persons sitting on counters
1,2 and 3 can serve 15, 25 and 30
customers per hour on average. Also
customers arrive with average rates of 12,
21 and 25 per hour in those three counters
respectively.
18 11/17/23

Example (M/M/1)
› If a customer wants to minimize her total time in
the system, which counter she should join?
› If a customer wants to minimize her total waiting
time in the system, which counter she should join?
› If a customer joins to a counter by seeing total
number of customers in that counter, which
counter she is likely to join?
› Which counter a counter person will prefer?
› What is the average total number of customers in
the system?
› In case the total number of customers in the
second counter cannot exceed 3 persons, what is
the revised service rate required assuming the
same arrival rate?
19 11/17/23

Example (M/M/1)

P0 W Wq L Lq r

Counter 1 0.2 0.333 0.267 4 3.2 0.8


Counter 2 0.16 0.25 0.21 5.25 4.41 0.84
Counter 3 0.167 0.2 0.167 5 4.167 0.83
20 11/17/23

You might also like