QUEUING THEORY | Waiting Line Models with Excel Applications

QUEUING THEORY
with Excel Applications
──────────────────────────────────────────────
A Comprehensive Learning Module
Operations Research | Waiting Line Models | Service Operations

Operations Research | Page 1

QUEUING THEORY | Waiting Line Models with Excel Applications

MODULE OVERVIEW

Module Title Queuing Theory (Waiting Line Models) with Excel Applications

Queuing System Components · M/M/1 Model · M/M/s (Multi-Server) Model ·

Topics Covered
M/M/1 with Finite Capacity · Cost Analysis · Excel Applications

Duration 8–10 hours (split into 3 sessions)

College students — Business, Engineering, IT, Hospitality, Operations

Target Students
Management

Prerequisites Basic probability, algebra, familiarity with MS Excel

Learning Objectives
By the end of this module, students will be able to:
• Identify and describe the components of a queuing (waiting line) system
• Distinguish among common queuing models using Kendall's Notation
• Compute key performance measures for the M/M/1 single-server model
• Compute key performance measures for the M/M/s multi-server model
• Apply the M/M/1 model with finite queue capacity
• Perform waiting line cost analysis to balance service cost and waiting cost
• Set up and solve queuing problems in Microsoft Excel using formulas
• Use Excel to determine the optimal number of servers that minimizes total cost

Operations Research | Page 2

QUEUING THEORY | Waiting Line Models with Excel Applications

PART 1: QUEUING SYSTEM FUNDAMENTALS

UNIT 1: Introduction to Queuing Theory

1.1 What is Queuing Theory?

Queuing Theory (also called Waiting Line Theory) is a branch of Operations Research that studies the
behavior of waiting lines. It helps managers decide on the right number of servers, service rates, or
facility layouts to balance customer waiting time against the cost of providing service.

📌 Why Queuing Theory Matters

Every business with customers (or items, calls, transactions) arriving for service faces a trade-
off:

• Too FEW servers → long waiting lines → frustrated customers → lost business

• Too MANY servers → idle staff → wasted labor cost

Queuing Theory provides the mathematical tools to find the BALANCE POINT —

the number of servers (or service rate) that minimizes TOTAL COST

(waiting cost + service cost).

1.2 Components of a Queuing System

ARRIVALS QUEUE (Waiting Line) SERVICE FACILITY DEPARTURES
● ● ● ──────► [ ][ ][ ][ ] ──────► ┌──────────┐ ──────► ✓ ✓ ✓
(λ = arrival rate) (customers waiting) │ Server 1 │ (served
│ Server 2 │ customers)
│ ... │
└──────────┘
(μ = service rate per server)

Component Description

Calling Population The source of customers/arrivals; can be infinite or finite

How customers arrive — usually modeled as a Poisson process with rate λ

Arrival Process
(lambda)

The rule for selecting the next customer to serve — most common: FCFS
Queue Discipline
(First-Come, First-Served)

Number of servers and the service time distribution — usually Exponential

Service Mechanism
with rate μ (mu)

Maximum number of customers allowed in the system (infinite or finite,

Queue Capacity
denoted N)

1.3 Kendall's Notation

Operations Research | Page 3
QUEUING THEORY | Waiting Line Models with Excel Applications
Queuing models are classified using Kendall's Notation: A / B / s / N / K

A / B / s / N / K

A = Arrival distribution (M = Poisson/Markovian, D = Deterministic, G = General)

B = Service time distribution (M = Exponential, D = Deterministic, G = General)
s = Number of servers (e.g., 1, 2, 3, ...)
N = System capacity (default = ∞ if omitted)
K = Calling population size (default = ∞ if omitted)

Example: M/M/1 means Poisson arrivals, Exponential service, 1 server, infinite capacity

1.4 Common Queuing Models in This Module

Model Description Typical Application

Single server, Poisson arrivals,

Single bank teller, single
M/M/1 Exponential service, infinite
checkout counter
capacity

Multiple (s) identical servers,

Bank with multiple tellers, call
M/M/s same arrival/service
center with multiple agents
assumptions

Single server with FINITE queue Small waiting area, limited

M/M/1/N
capacity N parking, limited seats

Operations Research | Page 4

QUEUING THEORY | Waiting Line Models with Excel Applications

UNIT 2: The M/M/1 Single-Server Model

2.1 Assumptions of the M/M/1 Model

• Arrivals follow a Poisson distribution with mean arrival rate λ (customers per unit time)
• Service times follow an Exponential distribution with mean service rate μ (customers per unit time)
• There is only ONE server
• Queue capacity is INFINITE (no limit on waiting line length)
• Calling population is INFINITE
• Queue discipline is FCFS (First-Come, First-Served)
• The system must be STABLE: λ < μ (utilization ratio ρ = λ/μ must be less than 1)

2.2 Key M/M/1 Formulas

Utilization Factor: ρ = λ / μ (must be < 1 for stability)

P(0) — Probability system is empty: P₀ = 1 − ρ

Lq — Average number waiting in queue: Lq = λ² / [μ(μ − λ)] = ρ² / (1 − ρ)

L — Average number in the system: L = λ / (μ − λ) = Lq + ρ

Wq — Average waiting time in queue: Wq = Lq / λ = λ / [μ(μ − λ)]

W — Average time in the system: W = L / λ = Wq + (1/μ)

Pn — Probability of exactly n customers in system: Pn = (1 − ρ) × ρⁿ

P(W>0) — Probability an arriving customer waits: = ρ

Find: ρ, P₀, L, Lq, W, Wq, and the probability of more than 3 customers in the bank.
Given data: λ = 12 customers/hour μ = 15 customers/hour

2.3 Worked Example — Bank Teller

📝 Problem: Single Bank Teller
A small rural bank in Antique has ONE teller window. Customers arrive at an average rate

of 12 customers per hour, following a Poisson distribution. The teller can process

customers at an average rate of 15 customers per hour, following an Exponential distribution.

Find: ρ, P₀, L, Lq, W, Wq, and the probability of more than 3 customers in the bank.

Given Data:
λ = 12 customers/hour μ = 15 customers/hour

Operations Research | Page 5

QUEUING THEORY | Waiting Line Models with Excel Applications
Step-by-Step Solution:
Check System Stability
STEP

1 ρ = λ/μ = 12/15 = 0.80

Since ρ = 0.80 < 1, the system IS STABLE (queue will not grow infinitely)

Compute P₀ — Probability the System is Empty

STEP

2 P₀ = 1 − ρ = 1 − 0.80 = 0.20

There is a 20% chance the bank has NO customers at any given time

Compute L — Average Number in the System

STEP

3 L = λ / (μ − λ) = 12 / (15 − 12) = 12 / 3 = 4.0 customers

On average, 4 customers are in the bank (waiting + being served)

Compute Lq — Average Number Waiting in Queue

STEP

4 Lq = λ² / [μ(μ − λ)] = 144 / [15(3)] = 144 / 45 = 3.2 customers

On average, 3.2 customers are waiting in line (not counting the one being served)

Compute W — Average Time in the System

STEP

5 W = L / λ = 4.0 / 12 = 0.333 hours = 20 minutes

On average, a customer spends 20 minutes total at the bank

Compute Wq — Average Waiting Time in Queue

STEP

6 Wq = Lq / λ = 3.2 / 12 = 0.267 hours = 16 minutes

On average, a customer waits 16 minutes BEFORE being served

Compute Probability of More Than 3 Customers

STEP P(n > 3) = 1 − [P₀ + P₁ + P₂ + P₃] = ρ⁴ (shortcut formula for M/M/1)
7 P(n > 3) = ρ⁴ = 0.80⁴ = 0.4096

There is a 40.96% chance more than 3 customers are in the bank at any time

2.4 Summary of Results

Measure Symbol Value Interpretation

Utilization ρ 0.80 Teller is busy 80% of the time

P(system
P₀ 0.20 20% chance of no customers
empty)

Avg. in system L 4.0 4 customers on average (waiting + served)

Avg. in queue Lq 3.2 3.2 customers waiting on average

Avg. time in
W 20 min Total time per customer
system

Avg. waiting
Wq 16 min Time before being served
time

Operations Research | Page 6

QUEUING THEORY | Waiting Line Models with Excel Applications

P(more than 3
P(n>3) 40.96% Significant chance of congestion
in system)

Operations Research | Page 7

QUEUING THEORY | Waiting Line Models with Excel Applications

UNIT 3: The M/M/s Multiple-Server Model

3.1 When to Use M/M/s

When MORE THAN ONE identical server is available (e.g., multiple bank tellers, multiple checkout
counters, multiple call center agents), we use the M/M/s model. All servers share ONE common queue.

3.2 Key M/M/s Formulas

Utilization Factor: ρ = λ / (s × μ) (must be < 1 for stability; s = number of servers)

P₀ — Probability system is empty:

P₀ = 1 / { Σₙ₌₀ˢ⁻¹ [(λ/μ)ⁿ / n!] + [(λ/μ)ˢ / s!] × [1 / (1−ρ)] }

Lq — Average number waiting in queue:

Lq = [P₀ × (λ/μ)ˢ × ρ] / [s! × (1−ρ)²]

L — Average number in the system: L = Lq + (λ/μ)

Wq — Average waiting time in queue: Wq = Lq / λ

W — Average time in the system: W = Wq + (1/μ)

💡 Important Note
The P₀ formula for M/M/s is complex to compute by hand because of the summation term.

This is exactly where EXCEL becomes essential — we will build a reusable Excel template

in Unit 5 that computes all M/M/s formulas automatically, even for large s values.

3.3 Worked Example — Multi-Teller Bank

📝 Problem: Bank with Multiple Tellers
The same bank now has TWO (s=2) teller windows, with the same arrival and service rates.

λ = 12 customers/hour μ = 15 customers/hour (per teller) s = 2 tellers

Find: ρ, P₀, L, Lq, W, Wq.

Check Stability
STEP

1 ρ = λ / (s×μ) = 12 / (2×15) = 12/30 = 0.40

Since ρ = 0.40 < 1, the system is stable with significant capacity to spare

STEP
Compute P₀ (the system-empty probability)
2 λ/μ = 12/15 = 0.80

Term 1: Σₙ₌₀¹ [(0.80)ⁿ/n!] = (0.80)⁰/0! + (0.80)¹/1! = 1 + 0.80 = 1.80

Term 2: [(0.80)²/2!] × [1/(1−0.40)] = [0.64/2] × [1/0.60] = 0.32 × 1.667 = 0.5333

P₀ = 1 / (1.80 + 0.5333) = 1 / 2.3333 = 0.4286

Operations Research | Page 8
QUEUING THEORY | Waiting Line Models with Excel Applications

Compute Lq
STEP Lq = [P₀ × (λ/μ)ˢ × ρ] / [s! × (1−ρ)²]
3 = [0.4286 × 0.64 × 0.40] / [2 × 0.36]

= 0.1098 / 0.72 = 0.1525 customers

Compute L, Wq, and W

STEP L = Lq + (λ/μ) = 0.1525 + 0.80 = 0.9525 customers
4 Wq = Lq/λ = 0.1525/12 = 0.0127 hours = 0.76 minutes

W = Wq + (1/μ) = 0.0127 + 0.0667 = 0.0794 hours = 4.77 minutes

3.4 Comparison: 1 Teller vs. 2 Tellers

Measure M/M/1 (1 Teller) M/M/2 (2 Tellers) Improvement

Utilization (ρ) 0.80 (80%) 0.40 (40%) Less congestion

Avg. in queue (Lq) 3.2 customers 0.15 customers 95% reduction

Avg. wait time (Wq) 16 minutes 0.76 minutes 95% reduction

Avg. time in system

20 minutes 4.77 minutes 76% reduction
(W)

✅ Key Insight
Adding just ONE more server dramatically reduces waiting time — from 16 minutes to under

1 minute! This demonstrates why queuing analysis is critical for staffing decisions.

However, the SECOND teller also costs money — this trade-off is explored in Unit 4.

Operations Research | Page 9

QUEUING THEORY | Waiting Line Models with Excel Applications

UNIT 4: Waiting Line Cost Analysis

4.1 The Cost Trade-off

The goal of queuing analysis is often to MINIMIZE TOTAL COST, which has two opposing components:

📌 Two Types of Cost

1. WAITING COST (Cw): Cost of customers' time spent waiting — lost goodwill, lost sales,

opportunity cost of customer time, or the cost of idle employees

waiting to be served (in internal queuing systems)

2. SERVICE COST (Cs): Cost of providing service — wages, equipment, facility costs

per server, per unit of time

As we ADD servers: Service Cost ↑ but Waiting Cost ↓

GOAL: Find the number of servers (s) that MINIMIZES Total Cost = Service Cost + Waiting Cost

4.2 Cost Formulas

Total Service Cost per hour = s × Cs (s = number of servers, Cs = cost per server per hour)

Total Waiting Cost per hour = L × Cw (L = avg. number in system, Cw = waiting cost per customer per hour)

Total Cost per hour (TC) = (s × Cs) + (L × Cw)

4.3 Worked Example — Optimal Number of Tellers

📝 Problem: Finding the Optimal Number of Tellers
Using the same bank: λ = 12/hr, μ = 15/hr per teller.

Cost of each teller (Cs) = ₱150 per hour (salary + benefits)

Cost of customer waiting (Cw) = ₱200 per hour per customer (lost goodwill/opportunity)

Find the OPTIMAL number of tellers (s) that minimizes Total Cost.

Solution — Compare s = 1, 2, and 3 Tellers:

L (avg. in Service Cost Waiting Cost
s (Tellers) ρ = λ/(sμ) Total Cost
system) (s×₱150) (L×₱200)

1 0.80 4.000 ₱150 ₱800.00 ₱950.00

2 0.40 0.9525 ₱300 ₱190.50 ₱490.50

3 0.267 0.812 ₱450 ₱162.40 ₱612.40

✅ Optimal Decision: s = 2 Tellers (Total Cost = ₱490.50/hour — LOWEST)

Operations Research | Page 10
QUEUING THEORY | Waiting Line Models with Excel Applications

With 1 teller: Total Cost = ₱950.00 (long waits dominate the cost)

With 2 tellers: Total Cost = ₱490.50 (best balance — MINIMUM cost)

With 3 tellers: Total Cost = ₱612.40 (extra server cost outweighs the small reduction in waiting)

CONCLUSION: The bank should staff 2 tellers — adding a 3rd teller is not cost-justified.

Operations Research | Page 11

QUEUING THEORY | Waiting Line Models with Excel Applications

PART 2: EXCEL APPLICATIONS

UNIT 5: Queuing Models in Microsoft Excel — Step-

by-Step

5.1 Overview of Excel Workbook

We will build a reusable Excel workbook with three sheets, allowing instant calculation of all queuing
measures for any λ, μ, and s values.

Excel Workbook Structure

Sheet 1 — M/M/1 Calculator: Single-server formulas with live input cells

Sheet 2 — M/M/s Calculator: Multi-server formulas including the complex P₀ summation

Sheet 3 — Cost Optimization: Table comparing total cost across different values of s

5.2 Sheet 1 — M/M/1 Calculator — Step by Step

Set up input cells

STEP B2: Label 'Arrival Rate (λ)' C2: Enter value, e.g., 12
1 B3: Label 'Service Rate (μ)' C3: Enter value, e.g., 15

These are the ONLY two cells you change to analyze different scenarios

Compute Utilization (ρ)

STEP B5: Label 'Utilization (ρ)'
2 C5: =C2/C3

Add conditional formatting: highlight RED if C5 ≥ 1 (unstable system warning)

Compute P₀ — Probability System is Empty

STEP

3 B6: Label 'P(0) - System Empty'

C6: =1-C5

Compute L — Average Number in System

STEP B7: Label 'L - Avg. Number in System'
4 C7: =C2/(C3-C2)

Use IFERROR wrapper to catch unstable systems: =IFERROR(C2/(C3-C2),"UNSTABLE")

Compute Lq — Average Number in Queue

STEP B8: Label 'Lq - Avg. Number in Queue'
5 C8: =(C2^2)/(C3*(C3-C2))

Alternative formula: =C7-C5 (since Lq = L − ρ)

Operations Research | Page 12

QUEUING THEORY | Waiting Line Models with Excel Applications

Compute W and Wq — Time Measures

B9: Label 'W - Avg. Time in System (hrs)' C9: =C7/C2
STEP

6 B10: Label 'Wq - Avg. Wait in Queue (hrs)' C10: =C8/C2

B11: Label 'W in minutes' C11: =C9*60

B12: Label 'Wq in minutes' C12: =C10*60

Compute Pn — Probability of Exactly n Customers

B14: Label 'n' | C14: Label 'P(n)'
STEP

7 B15:B20: Enter n = 0,1,2,3,4,5

C15: =(1-$C$5)*$C$5^B15

Copy down C15:C20 to get the full probability distribution

Complete M/M/1 Excel Formula Reference:

Cell(s) Label Formula / Entry Notes

C2 Arrival rate λ 12 User input

C3 Service rate μ 15 User input

Must be < 1 for a valid/stable

C5 Utilization ρ =C2/C3
model

Probability of zero customers

C6 P₀ =1-C5
in system

Average number of
C7 L =C2/(C3-C2)
customers in the system

Average number of
C8 Lq =(C2^2)/(C3*(C3-C2))
customers waiting

C9 W (hours) =C7/C2 Average time in system

Average time waiting in

C10 Wq (hours) =C8/C2
queue

Convert hours to minutes for

C11 W (minutes) =C9*60
clarity

C12 Wq (minutes) =C10*60 Convert hours to minutes

Probability formula; copy

C15 P(n=0) =(1-$C$5)*$C$5^B15
down for n=1,2,3...

Probability of more than k

C22 P(n>k) shortcut =C5^(B22+1)
customers in system

5.3 Sheet 2 — M/M/s Calculator — Step by Step

The M/M/s model requires a summation term, which Excel handles elegantly using helper cells and the
SUMPRODUCT or SERIESSUM function.

Set up inputs
STEP B2: 'Arrival Rate (λ)' C2: 12
1 B3: 'Service Rate (μ)' C3: 15

B4: 'Number of Servers (s)' C4: 2

Operations Research | Page 13

QUEUING THEORY | Waiting Line Models with Excel Applications

Compute ρ (utilization) and λ/μ

STEP B6: 'λ/μ ratio' C6: =C2/C3
2 B7: 'Utilization ρ' C7: =C2/(C4*C3)

Conditional formatting: flag RED if C7 ≥ 1

Build the summation helper table for P₀

In a helper area (e.g., E2:F10), list n = 0 to (s−1)
STEP E2:E_(s-1): n values 0, 1, 2, ... up to s−1
3 F2: =($C$6^E2)/FACT(E2) (computes (λ/μ)ⁿ / n! for each n)

Copy down F2 for all n from 0 to s−1

G2: =SUM(F2:F_last) (this is the summation term, Σ)

Compute P₀ using the full formula

H2: 'P0 second term'
STEP

4 H2: =(($C$6^$C$4)/FACT($C$4))*(1/(1-$C$7))

C8: 'P0' C8: =1/(G2+H2)

This combines the summation (G2) and the last term (H2) into the full P₀ formula

Compute Lq, L, Wq, W

C9: 'Lq' C9: =(C8*(C6^C4)*C7)/(FACT(C4)*(1-C7)^2)
STEP

5 C10: 'L' C10: =C9+C6

C11: 'Wq' C11: =C9/C2

C12: 'W' C12: =C11+(1/C3)

Complete M/M/s Excel Formula Reference:

Cell(s) Label Formula / Entry Notes

C2 Arrival rate λ 12 User input

Service rate μ (per

C3 15 User input
server)

User input — try different

C4 Number of servers s 2
values

Used repeatedly in
C6 λ/μ ratio =C2/C3
subsequent formulas

C7 Utilization ρ =C2/(C4*C3) Must be < 1

E2:E9 Helper: n values 0,1,2,...,s−1 List integers from 0 to s−1

F2 Helper: (λ/μ)ⁿ/n! =($C$6^E2)/FACT(E2) Copy down for each n

Adjust range to match s−1

G2 Summation total =SUM(F2:F9)
rows used

Last term of P0 =(($C$6^$C$4)/FACT($C$4))*(1/(1-

H2
formula $C$7))

Combines summation + last

C8 P0 =1/(G2+H2)
term

C9 Lq =(C8*(C6^C4)*C7)/(FACT(C4)*(1- Average number waiting

Operations Research | Page 14

QUEUING THEORY | Waiting Line Models with Excel Applications

C7)^2)

C10 L =C9+C6 Average number in system

C11 Wq =C9/C2 Average wait time in queue

C12 W =C11+(1/C3) Average time in system

5.4 Sheet 3 — Cost Optimization Table

Build a table that automatically computes Total Cost for s = 1, 2, 3, 4, 5 servers and highlights the
optimal choice.

Set up cost inputs

B2: 'Service cost per server/hr (Cs)' C2: 150
STEP

1 B3: 'Waiting cost per customer/hr (Cw)' C3: 200

B4: 'Arrival rate λ' C4: 12

B5: 'Service rate μ' C5: 15

Build the comparison table

Row 8: Headers: s | ρ | L | Service Cost | Waiting Cost | Total Cost

Column A9:A13: s = 1, 2, 3, 4, 5
STEP B9: =C4/(A9*C5) (utilization for each s)
2 C9: Use the M/M/s L formula (link to Sheet 2 calculations, or rebuild inline)

D9: =A9*$C$2 (service cost)

E9: =C9*$C$3 (waiting cost)

F9: =D9+E9 (total cost)

Highlight the optimal solution

Select F9:F13 (Total Cost column)
STEP

3 Home → Conditional Formatting → Highlight Cell Rules → Equal To

Set the formula: =MIN($F$9:$F$13)

Choose a green fill — this automatically highlights the lowest-cost row

Use Excel Solver to confirm (optional)

Data tab → Solver
STEP Set Objective: the Total Cost cell for a chosen s formula cell
4 By Changing Variable Cells: the 's' input cell

Add constraint: s = integer, s ≥ 1, ρ < 1

Solving method: Evolutionary (since the model is non-linear and uses factorial functions)

Cost Optimization Table Layout (Final Result):

s ρ L Service Cost Waiting Cost Total Cost

1 0.800 4.000 ₱150 ₱800.00 ₱950.00

Operations Research | Page 15

QUEUING THEORY | Waiting Line Models with Excel Applications

2 0.400 0.953 ₱300 ₱190.50 ₱490.50

3 0.267 0.812 ₱450 ₱162.40 ₱612.40

4 0.200 0.804 ₱600 ₱160.80 ₱760.80

5 0.160 0.801 ₱750 ₱160.13 ₱910.13

✅ Excel confirms: s = 2 is optimal (Total Cost = ₱490.50, lowest in the table)

Operations Research | Page 16

QUEUING THEORY | Waiting Line Models with Excel Applications

SEATWORK 1: M/M/1 Single-Server Model

📋 Instructions
Show all formulas and calculations clearly. Round answers to 4 decimal places where
applicable.

Time allowed: 25 minutes.

Problem — Sari-Sari Store Checkout

A sari-sari store in San Jose, Antique has one checkout counter. Customers arrive at an average rate of
20 customers per hour following a Poisson distribution. The cashier can serve customers at an average
rate of 25 customers per hour following an Exponential distribution.

(a) Check system stability and compute ρ: (4 pts)

ρ = λ/μ = ___ / ___ = _______________

Is the system stable? (Yes/No): _______________

(b) Compute P₀ — probability the system is empty: (4 pts)

P₀ = 1 − ρ = 1 − ___ = _______________

(c) Compute L — average number of customers in the system: (4 pts)

L = λ / (μ−λ) = ___ / (___ − ___) = _______________

(d) Compute Lq — average number of customers in the queue: (4 pts)

Lq = λ² / [μ(μ−λ)] = ___² / [___ × (___ − ___)] = _______________

(e) Compute W and Wq (in minutes): (6 pts)

W = L / λ = ___ / ___ = _______________ hours = _______________ minutes

Wq = Lq / λ = ___ / ___ = _______________ hours = _______________ minutes

(f) Find the probability that more than 2 customers are in the store: (4 pts)
P(n>2) = ρ³ = (___)³ = _______________

(g) Excel Application: Set up this problem in Excel using the formulas taught. Record the cell formulas
you used: (4 pts)
Formula for ρ: =_______________

Formula for L: =_______________

Formula for Wq: =_______________

Operations Research | Page 17

QUEUING THEORY | Waiting Line Models with Excel Applications

SEATWORK 2: M/M/s Multi-Server Model

📋 Instructions
Show all calculations including the P₀ summation term. Use Excel for verification.

Time allowed: 35 minutes.

Problem — Cooperative Loan Processing Center

A multi-purpose cooperative has 3 loan officers (s=3) processing member loan applications. Applications
arrive at an average rate of 18 per hour, and each loan officer can process applications at an average
rate of 8 per hour.

(a) Compute λ/μ and ρ: (6 pts)

λ/μ = __18_ / _8_ = _______________2.25

ρ = λ/(s×μ) = __18_ / (_3__×__8_) = _______________0.75

Is the system stable? (Yes/No): _______________Stable

(b) Compute P₀ using the summation formula. Show your work for each term: (10 pts)
Term for n=0: (λ/μ)⁰/0! = _______________ 1

Term for n=1: (λ/μ)¹/1! = _______________2.25

Term for n=2: (λ/μ)²/2! = _______________2.53125

Summation Σ (n=0 to s−1) = _______________5.78125

Last term: [(λ/μ)ˢ/s!] × [1/(1−ρ)] = (2.25)^3 x 1/0.25 = 1.8984X4 = 7.5936

P₀ = 1 / (Summation + Last term) = 1/5.7812+7.5936 =1/13.3748= 0.0747

(c) Compute Lq, L, Wq, and W: (12 pts)

Lq = [P₀ × (λ/μ)ˢ × ρ] / [s! × (1−ρ)²] = 0.0747 x 11.3906 x 0.75/ (6 x 0.0625) = 0.6381/0.375 =
1.7016

L = Lq + (λ/μ) = 1.7016+2.25 = 3.9516

Wq = Lq / λ = 1.7016/18= 0.0953 hours = 5.72 minutes

W = Wq + (1/μ) = 0.0953+0.125=0.2203 hours = 13.22 minutes

(d) Excel Application: Build the M/M/s calculator in Excel as taught. (8 pts)
Helper table formula for (λ/μ)ⁿ/n!: =_______________

Formula for P₀: =_______________

Formula for Lq: =_______________

Excel computed values — P₀=_______ Lq=_______ L=_______ Wq=_______ W=_______

Do your manual calculations match Excel? (Yes/No): _______________

Operations Research | Page 18

QUEUING THEORY | Waiting Line Models with Excel Applications

SEATWORK 3: Waiting Line Cost Analysis

📋 Instructions
Build the complete cost comparison table. Use Excel as taught to confirm your answer.

Time allowed: 30 minutes.

Problem — Government Service Office (LTO-style Window)

A government licensing office is deciding how many service windows (s) to open. Customers arrive at a
rate of 24 per hour. Each window can process 10 customers per hour. The cost of staffing each window
is ₱120/hour. The cost of customer waiting (estimated lost goodwill/productivity) is ₱150 per customer
per hour.

Complete the cost comparison table for s = 3, 4, and 5 windows: (24 pts)

Service Cost (s × Waiting Cost (L × Total

s ρ = λ/(sμ) L (use M/M/s)
₱120) ₱150) Cost

3 0.80 5.048 360 757.20 1,117.20

4 0.60 2.829 480 424.35 904.35

5 0.48 2.524 600 378.60 978.60

Optimal number of windows: s = 4 Minimum Total Cost = ₱904.35

Excel Application: (6 pts)

Formula for Total Cost cell (general form): =_______________ + _______________

Which conditional formatting rule did you use to highlight the minimum cost row?

Formula: =_______________

Did Excel Solver's recommended number of windows match your manual answer? (Yes/No):
_______________

Operations Research | Page 19

QUEUING THEORY | Waiting Line Models with Excel Applications

UNIT 6: Practice Problems

Practice Set A — M/M/1 Model

A1. Internet Café Service

An internet café has one service counter for printing and encoding requests. Customers arrive at a rate
of 15 per hour. The attendant can serve 20 customers per hour. Find ρ, P₀, L, Lq, W, and Wq.

A2. Single ATM Machine

A single ATM machine experiences customer arrivals at a rate of 10 per hour. Average service time is 4
minutes per transaction. Convert to compatible units, then find all M/M/1 performance measures. What is
the probability a customer will have to wait?

A3. Probability Distribution

For an M/M/1 system with λ=8/hr and μ=10/hr, compute P(0), P(1), P(2), P(3), P(4), and P(n>4). Verify
that ΣP(n) for n=0 to 4 plus P(n>4) equals 1.

Practice Set B — M/M/s Model

B1. Two-Counter Service Station

A vehicle inspection station has 2 inspection lanes. Vehicles arrive at 16 per hour. Each lane can inspect
10 vehicles per hour. Find ρ, P₀, Lq, L, Wq, and W. Compare with what would happen if there were only
1 lane.

B2. Call Center Staffing

A BPO call center receives calls at a rate of 40 per hour. Each agent can handle 12 calls per hour.
Currently the center has 4 agents. (a) Is the system stable? (b) Find all M/M/s performance measures.
(c) What is the minimum number of agents needed for a stable system?

B3. Excel Sensitivity Analysis

Using Problem B2's call center, build an Excel table testing s = 4, 5, 6, and 7 agents. Show how Wq
decreases as more agents are added. At what point does adding agents provide only marginal
improvement (diminishing returns)?

Practice Set C — Cost Optimization

C1. Hospital Triage Desk

A hospital is deciding on the number of triage nurses. Patients arrive at 14 per hour. Each nurse can
triage 6 patients per hour. Nurse cost = ₱300/hour. Patient waiting cost (health risk consideration) =
₱500/hour per patient. Find the optimal number of nurses (s=3,4,5) that minimizes total cost.

C2. Drive-Thru Optimization

A fast food drive-thru is considering adding a second order window. Current single window: λ=30/hr,
μ=35/hr. Cost per window = ₱200/hour. Cost of customer waiting (lost sales/goodwill) = ₱100/hour per
customer. Should the restaurant add the second window? Show full cost comparison.

Operations Research | Page 20

QUEUING THEORY | Waiting Line Models with Excel Applications

C3. Comprehensive Excel Model

Build a complete Excel workbook combining M/M/s calculations AND cost optimization for the Drive-Thru
problem (C2). Include: input cells, M/M/s formulas with summation, a cost comparison table for s=1 to 4,
and conditional formatting highlighting the optimal choice. Use Excel Solver to confirm your answer.

Operations Research | Page 21

QUEUING THEORY | Waiting Line Models with Excel Applications

ANSWER KEY & MODEL SOLUTIONS

⚠️ Instructor Reference
Model solutions for all seatworks and key practice problems.

Encourage students to verify manual calculations against their Excel models.

Seatwork 1 — Sari-Sari Store Checkout

Measure Calculation Answer

ρ 20/25 0.80

P₀ 1−0.80 0.20

L 20/(25−20) = 20/5 4.0 customers

Lq 400/[25×5] = 400/125 3.2 customers

W 4.0/20 = 0.20 hrs 12 minutes

Wq 3.2/20 = 0.16 hrs 9.6 minutes

P(n>2) 0.80³ 0.512 (51.2%)

Seatwork 2 — Cooperative Loan Processing

λ=18, μ=8, s=3. λ/μ = 2.25. ρ = 18/(3×8) = 18/24 = 0.75
Summation terms: n=0: 1.000, n=1: 2.250, n=2: 2.531 → Σ = 5.781
Last term: (2.25³/6) × (1/0.25) = (11.39/6) × 4 = 1.898 × 4 = 7.594
P₀ = 1/(5.781+7.594) = 1/13.375 = 0.0748
Lq = (0.0748 × 11.39 × 0.75)/(6 × 0.0625) = 0.6391/0.375 = 1.704 customers
L = 1.704 + 2.25 = 3.954 customers
Wq = 1.704/18 = 0.0947 hrs = 5.68 minutes
W = 0.0947 + 0.125 = 0.2197 hrs = 13.18 minutes

Seatwork 3 — Government Service Office

λ=24, μ=10, Cs=₱120, Cw=₱150
Total
s ρ L (approx.) Service Cost Waiting Cost
Cost

3 0.80 4.500 ₱360 ₱675.00 ₱1,035.00

4 0.60 2.617 ₱480 ₱392.55 ₱872.55

5 0.48 2.485 ₱600 ₱372.75 ₱972.75

Optimal: s = 4 windows (Total Cost = ₱872.55/hour — lowest)

Practice Set Key Results

Problem Key Result

ρ=0.75, P₀=0.25, L=3.0, Lq=2.25, W=12min,

A1 (Internet Café)
Wq=9min

Operations Research | Page 22

QUEUING THEORY | Waiting Line Models with Excel Applications

μ=15/hr (4 min=1/15 hr), ρ=0.667, P(wait)=0.667,

A2 (ATM)
Wq=8.89min

2 lanes: ρ=0.80, Lq=2.844; 1 lane would be

B1 (Vehicle Inspection)
UNSTABLE (ρ=1.6>1)

ρ=0.833 (4 agents), stable; minimum agents

B2 (Call Center)
needed=4 (since λ/μ=3.33)

s=4 optimal; compare costs at s=3 (₱2,800ish) vs

C1 (Hospital Triage)
s=4 (lower) vs s=5

2nd window reduces Wq significantly; compare

C2 (Drive-Thru)
total cost to justify addition

Glossary of Key Queuing Theory Terms

Term Definition

Arrival Rate (λ) The average number of customers/units arriving per unit of time

Calling Population The source of arrivals; can be finite or infinite

FCFS First-Come, First-Served; the most common queue discipline

Kendall's Notation Classification system for queuing models: A/B/s/N/K

Average number of customers in the entire system (queue + being

L
served)

Lq Average number of customers waiting in the queue (not being served)

Single-server queuing model with Poisson arrivals and Exponential

M/M/1
service

M/M/s Multi-server queuing model with s identical servers

Poisson Distribution Probability distribution commonly used to model random arrivals

P₀ Probability that the system has zero customers (idle system)

Pn Probability of exactly n customers in the system

Queue Discipline The rule used to determine which customer is served next

The average number of customers a single server can serve per unit
Service Rate (μ)
time

Stability Condition Requirement that ρ < 1 (λ < sμ) for the queue not to grow infinitely

The fraction of time the server(s) are busy; ρ = λ/μ (single) or λ/(sμ)
Utilization Factor (ρ)
(multi)

Average time a customer spends in the entire system (waiting +

W
service)

Average time a customer spends waiting in the queue before service

Wq
begins

The cost associated with customers waiting (lost goodwill, productivity,

Waiting Cost
sales)

The cost of providing service capacity (staff wages, equipment,

Service Cost
facilities)

Operations Research | Page 23