Project Management — All Math Types Explained URP 3291 | KUET

All Math Types — Fully Explained

Project Management Calculations with Worked Examples

URP 3291 | Operations Research & Systems Analysis

Khulna University of Engineering & Technology
Chapter 18: Project Management

Type 1 Type 2 Type 3 Type 4 Type 5

Path & Critical Path ES / EF / LS / LF / Slack te & Variance z-score & Probability Crashing

HOW TO USE THIS PDF

Each section covers one calculation type with: concept explanation,
formula box, step-by-step procedure, and a fully worked example.
All examples are from the Stevenson textbook Chapter 18.

Operations Research & Systems Analysis | Page 1

Project Management — All Math Types Explained URP 3291 | KUET

Path & Critical Path

1 Finding which sequence of activities governs project duration

Core idea: A network has multiple routes from start to finish. Each route is called a path. The path that takes the
longest time controls when the project finishes — that is the critical path.

Path length = sum of all activity times on that path

Critical path = the path with the LARGEST length
Project duration = length of the critical path
Slack of any path = Critical path length - That path length

KEY RULE: The critical path always has Slack = 0.

Any other path with Slack > 0 can be delayed by that amount
without delaying the entire project.

Step-by-step procedure:

1 Draw or read the network diagram carefully.

2 List ALL paths from the first node to the last node.

3 Add up the activity times along each path to get its total length.

4 The longest total = critical path = project duration.

5 Slack for each path = Critical path length minus that path length.

6 State which activities are critical (those on the zero-slack path).

Worked example — Bank project (Textbook Example 1):

Network activities and times (weeks): Locate facilities (1-2)=8, Interview staff (1-3)=4, Order furniture (2-4)=6,
Remodel (2-5)=11, Hire & train (3-5)=9, Furniture set-up (4-5)=3, Move in (5-6)=1.

Length
Path Activities in path Calculation Slack Status
(wks)

Locate, Order furn, Set-up, Move

1-2-4-5-6 8+6+3+1 18 20-18 = 2 Non-critical
in

CRITICAL
1-2-5-6 Locate, Remodel, Move in 8+11+1 20 20-20 = 0
PATH

1-3-5-6 Interview, Hire & train, Move in 4+9+1 14 20-14 = 6 Non-critical

ANSWER: Critical path = 1-2-5-6 | Project duration = 20 weeks

Critical activities: 1-2 (Locate), 2-5 (Remodel), 5-6 (Move in) — cannot be delayed at all.
Activity 2-4 and 4-5 have 2 weeks slack each (shared). Activity 1-3 and 3-5 share 6 weeks.

Operations Research & Systems Analysis | Page 2

Project Management — All Math Types Explained URP 3291 | KUET

Computing Algorithm: ES, EF, LS, LF & Slack

2 Forward pass and backward pass through the network

This method finds four values for every individual activity, then uses them to identify the critical path and exactly
how much slack each activity has — more precise than the path-listing method.

Symbol Full name Plain meaning Formula

Earliest time an activity CAN begin (all predecessors

ES Earliest Start Largest EF of all predecessors
finished)

EF Earliest Finish Earliest time an activity CAN finish ES + t

LS Latest Start Latest it can start WITHOUT delaying the project LF - t

LF Latest Finish Latest it can finish WITHOUT delaying the project Smallest LS of all successors

Slack = LS - ES (identical result: LF - EF)

Activities where Slack = 0 --> they are ON the CRITICAL PATH

FORWARD PASS (left to right) — finding ES and EF:

1 All starting activities get ES = 0.

2 EF = ES + t for every activity.

3 The next activity's ES = EF of the one just before it.

4 MERGING RULE: if two or more arrows arrive at the same node, the leaving activity's ES = LARGEST of all incoming EF values (i

5 The EF of the very last activity = Project Duration.

BACKWARD PASS (right to left) — finding LF and LS:

1 Last activity: LF = EF = project duration.

2 LS = LF - t for every activity.

3 The preceding activity's LF = LS of the one just after it.

4 SPLITTING RULE: if one node leads to two or more arrows, the incoming activities' LF = SMALLEST of all outgoing LS values (mu

WHY largest EF on forward pass?

An activity cannot start until every single predecessor is finished.
If A finishes at week 8 and B finishes at week 14, C must wait until week 14.

WHY smallest LS on backward pass?

An activity must finish in time for the EARLIEST of its successors.
If D must start by week 10 and E must start by week 16, the predecessor
must finish by week 10 (the tighter constraint).

Fully worked example — Bank project (same network as Type 1):

Activity times: 1-2=8, 1-3=4, 2-4=6, 2-5=11, 3-5=9, 4-5=3, 5-6=1. Project duration = 20 weeks.

Activity t ES EF=ES+t LF LS=LF-t Slack=LS-ES Critical?

1-2 8 0 8 8 0 0 YES

Operations Research & Systems Analysis | Page 3

Project Management — All Math Types Explained URP 3291 | KUET

1-3 4 0 4 10 6 6 no

2-4 6 8 14 16 10 2 no

2-5 11 8 19 19 8 0 YES

3-5 9 4 13 19 10 6 no

4-5 3 14 17 19 16 2 no

5-6 1 19 20 20 19 0 YES

HOW TO FIND ES for activity 5-6:

Node 5 is reached from activities 2-5 (EF=19), 3-5 (EF=13), 4-5 (EF=17).
Take the LARGEST: ES(5-6) = 19.

HOW TO FIND LF for activity 1-2:

Node 2 leads out to activities 2-4 (LS=10) and 2-5 (LS=8).
Take the SMALLEST: LF(1-2) = 8.

Critical path = 1-2-5-6 (activities with Slack = 0)

Project duration = 20 weeks (EF of last activity 5-6)
Verify: LS-ES = LF-EF for every row (both columns must give the same slack).

Operations Research & Systems Analysis | Page 4

Project Management — All Math Types Explained URP 3291 | KUET

Expected Time (te) and Variance

3 Converting three estimates into one expected value and a spread measure

When activity times are uncertain, PERT requires three estimates per activity. These are converted into a single
expected time (te) and a variance (sigma2) before any further calculation.

Symbol Name Meaning

to Optimistic time Time needed under BEST possible conditions — everything goes right

tm Most likely time Most PROBABLE time — normal working conditions

tp Pessimistic time Time needed under WORST conditions — everything goes wrong

te = (to + 4 * tm + tp) / 6
Variance (sigma^2) = (tp - to)^2 / 36
Path mean = sum of te for ALL activities on that path
Path variance = sum of sigma^2 for ALL activities on that path
Path standard deviation (sigma) = sqrt (path variance)

WHY is tm multiplied by 4?
Because tm is the most probable time, it carries more weight in the average.
The beta distribution assigns 4x weight to the most likely estimate.
WHY divide by 36 for variance?
The standard deviation is (tp-to)/6, so variance = [(tp-to)/6]^2 = (tp-to)^2/36.

Worked example — Textbook Example 5 (three paths):

Path Activity to tm tp te = (to+4tm+tp)/6 sigma^2 = (tp-to)^2/36

a-b-c a 1 3 4 (1+12+4)/6 = 17/6 = 2.83 (4-1)^2/36 = 9/36 = 0.25

b 2 4 6 (2+16+6)/6 = 24/6 = 4.00 (6-2)^2/36 = 16/36 = 0.44

c 2 3 5 (2+12+5)/6 = 19/6 = 3.17 (5-2)^2/36 = 9/36 = 0.25

PATH 0.25+0.44+0.25=0.94
— — — — 2.83+4.00+3.17 = 10.00 wks
TOTAL sigma=0.97

d-e-f d 3 4 5 (3+16+5)/6 = 24/6 = 4.00 (5-3)^2/36 = 4/36 = 0.11

e 3 5 7 (3+20+7)/6 = 30/6 = 5.00 (7-3)^2/36 = 16/36 = 0.44

f 5 7 9 (5+28+9)/6 = 42/6 = 7.00 (9-5)^2/36 = 16/36 = 0.44

PATH
0.11+0.44+0.44=0.99
TOTAL — — — — 4.00+5.00+7.00 = 16.00 wks CRITICAL
sigma=1.00
*CRITICAL*

g-h-i g 2 3 6 (2+12+6)/6 = 20/6 = 3.33 (6-2)^2/36 = 16/36 = 0.44

h 4 6 8 (4+24+8)/6 = 36/6 = 6.00 (8-4)^2/36 = 16/36 = 0.44

i 3 4 6 (3+16+6)/6 = 25/6 = 4.17 (6-3)^2/36 = 9/36 = 0.25

PATH 0.44+0.44+0.25=1.13
— — — — 3.33+6.00+4.17 = 13.50 wks
TOTAL sigma=1.07

Operations Research & Systems Analysis | Page 5

Project Management — All Math Types Explained URP 3291 | KUET

Critical path = d-e-f (largest path mean = 16.00 weeks)

Path means: a-b-c=10.00, d-e-f=16.00 (critical), g-h-i=13.50
Path sigmas: a-b-c=0.97, d-e-f=1.00, g-h-i=1.07
These sigmas are used in Type 4 to compute completion probabilities.

Operations Research & Systems Analysis | Page 6

Project Management — All Math Types Explained URP 3291 | KUET

z-score and Completion Probability

4 Using path means and sigmas to find the chance of finishing by a deadline

Once you have the path mean and sigma from Type 3, you can calculate the probability that each path — and
therefore the whole project — finishes by any given deadline T.

z = (T - path mean) / path sigma

P(path done by T) = area to the LEFT of z on the normal curve
P(project done by T) = P(path 1) x P(path 2) x P(path 3) x ...
P(NOT done by T) = 1 - P(done by T)

CRITICAL RULE: If z >= +2.50, treat P = 1.000 (the path is almost certain to finish on time).
Positive z = deadline is AFTER expected finish = P > 50%.
Negative z = deadline is BEFORE expected finish = P < 50%.
Paths are independent when no activity appears on more than one path.

Key z-values to memorise:

z value P (area to left) Interpretation

-2.00 0.0228 Only 2.3% chance of finishing by deadline

-1.00 0.1587 15.9% chance — deadline is 1 sigma too early

0.00 0.5000 50% chance — deadline equals expected finish

+1.00 0.8413 84.1% chance — very common exam answer

+1.40 0.9192 91.9% chance

+1.67 0.9525 95.3% chance

+2.00 0.9772 97.7% chance

+2.50 or more 1.0000 Use rule of thumb: treat as 100%

Worked example A — Target T = 17 weeks:

Using path data from Type 3: a-b-c mean=10, sigma=0.97 | d-e-f mean=16, sigma=1.00 | g-h-i mean=13.5,
sigma=1.07

Path Mean Sigma z = (17-mean)/sigma P from table

a-b-c 10.00 0.97 (17-10)/0.97 = 7.0/0.97 = +7.22 z>=2.50 --> 1.000

d-e-f (critical) 16.00 1.00 (17-16)/1.00 = 1.0/1.00 = +1.00 z=+1.00 --> 0.8413

g-h-i 13.50 1.07 (17-13.5)/1.07 = 3.5/1.07 = +3.27 z>=2.50 --> 1.000

P(project done by week 17) = 1.000 x 0.8413 x 1.000 = 0.8413 = 84.13%

Only the critical path (d-e-f) mattered here because the others had z >= 2.50.

Worked example B — Target T = 15 weeks (trickier — negative z):

Path Mean Sigma z = (15-mean)/sigma P from table

a-b-c 10.00 0.97 (15-10)/0.97 = 5.0/0.97 = +5.15 z>=2.50 --> 1.0000

Operations Research & Systems Analysis | Page 7

Project Management — All Math Types Explained URP 3291 | KUET

d-e-f 16.00 1.00 (15-16)/1.00 = -1.0/1.00 = -1.00 -1.00 --> 0.1587

g-h-i 13.50 1.07 (15-13.5)/1.07 = 1.5/1.07 = +1.40 +1.40 --> 0.9192

NOTE on negative z = -1.00 for path d-e-f:

Target (15 wks) is EARLIER than expected finish (16 wks).
This means it is unlikely to finish on time -- P is less than 50%.
From the normal table: area left of z=-1.00 = 0.1587 (only 15.9% chance).

P(project done by week 15) = 1.0000 x 0.1587 x 0.9192 = 0.1459 = 14.59%

P(project NOT done by week 15) = 1 - 0.1459 = 0.8541 = 85.41%
Interpretation: only a 14.6% chance of finishing within 15 weeks.

Operations Research & Systems Analysis | Page 8

Project Management — All Math Types Explained URP 3291 | KUET

Crashing — Time-Cost Trade-off

5 Shortening the project by spending more money on critical activities

What is crashing? Adding extra resources (money, overtime, better equipment) to shorten specific activity durations.
This increases direct costs but reduces indirect costs (overhead, facilities, supervision). The goal is to minimise
total cost.

Net saving per crash unit = Indirect cost/day - Crash cost/day

STOP crashing when: Crash cost/day > Indirect cost/day
(i.e. when it costs more to crash than the saving from reduced project duration)

KEY RULE: Only crash CRITICAL PATH activities.

Crashing a non-critical activity saves zero time on the project.
KEY RULE: Crash the CHEAPEST eligible critical activity first.
Sort by crash cost per unit and always pick the lowest.
KEY RULE: When TWO paths become critical simultaneously,
you must shorten BOTH. Compare: (a) one shared activity on both paths,
vs (b) cheapest on each path separately. Pick the lower total cost.

Step-by-step crashing procedure:

1 Find all paths. Identify the critical path and project duration.

2 List only the critical path activities. Rank them by crash cost/day (cheapest first).

3 Crash the cheapest by 1 day. Subtract 1 from its normal time. Recalculate path lengths.

4 Check: has any other path length now equalled the (shortened) critical path? If yes, both paths are now critical.

5 If ONE path is still critical: repeat from Step 2 on the new critical path.

6 If TWO paths are critical: you need to shorten BOTH simultaneously. Find cheapest option (shared activity or one from each path).

7 After each crash: check if savings still exceed crash cost. If NOT: STOP.

Worked example — Textbook Example 7:

Indirect costs = $1,000 per day. Two paths: a-b-f = 18 days | c-d-e-f = 20 days (critical).

Critical path activity Crash cost/day Max days can crash

c $300 1

e $600 2

d $700 3

f $800 1

a (not on crit. path) n/a cannot crash

b (not on crit. path) $500 only if b-path becomes critical

Roun Crash Indirect Net Path

Action taken Path a-b-f
d cost saving saving c-d-e-f

Start No crashing — — — 18 days 20 days

1 Crash c by 1 day (cheapest at $300) $300 $1,000 $700 18 19

Operations Research & Systems Analysis | Page 9

Project Management — All Math Types Explained URP 3291 | KUET

18 BOTH
2 Crash e by 1 day ($600, c exhausted) $600 $1,000 $400 18
critical now!

Crash f by 1 day ($800) — f is on BOTH

3 $800 $1,000 $200 17 17
paths!

4 Options: b($500)+e($600)=$1,100 > STOP

$1,100 $1,000 -$100 loss —
STOP $1,000 saving. No profit. HERE

WHY crash f in round 3 instead of b+e separately?

Both paths are now critical (both = 18 days).
Option A: crash activity f (on both paths) for $800 --> shortens both simultaneously.
Option B: crash b on path a-b-f ($500) + crash e on path c-d-e-f ($600) = $1,100.
$800 < $1,100 --> Option A (crash f) is cheaper. Always compare before deciding.

Optimal crashing plan: crash c (1 day, $300) + e (1 day, $600) + f (1 day, $800)
Total crashing cost = $1,700 | Indirect saving = 3 days x $1,000 = $3,000
Net saving = $3,000 - $1,700 = $1,300 | Final project duration = 17 days

Operations Research & Systems Analysis | Page 10

Project Management — All Math Types Explained URP 3291 | KUET

Complete Formula Sheet & Quick Reference

TYPE 1 — PATH: Path length = sum of activity times on that path

Slack = Critical path length - Path length

TYPE 2 — ALGORITHM: EF = ES + t LS = LF - t
Slack = LS - ES (same as LF - EF)
Forward merge: ES = LARGEST EF of predecessors
Backward split: LF = SMALLEST LS of successors

TYPE 3 — ESTIMATES: te = (to + 4 * tm + tp) / 6

Variance (sigma^2) = (tp - to)^2 / 36
Path sigma = sqrt(sum of variances on path)

TYPE 4 — PROBABILITY: z = (T - path mean) / path sigma

P(project) = P(path1) x P(path2) x P(path3)
If z >= 2.50: use P = 1.000

TYPE 5 — CRASHING: Net saving = Indirect cost - Crash cost (per unit)
STOP when Crash cost > Indirect saving

Critical rules — never get these wrong in the exam:

Rule What to do Why

Activity cannot start until ALL predecessors

Forward pass — merging Take LARGEST EF of all incoming activities
are done

Must finish before the EARLIEST successor

Backward pass — splitting Take SMALLEST LS of all outgoing activities
needs to start

Rule of thumb — virtually certain to finish on

z >= +2.50 Set P = 1.000 regardless of actual table value
time

Negative z P < 0.50 (look up left tail of normal table) Deadline is earlier than expected finish

Non-critical activities do not control project

Crashing target Only crash activities on the CRITICAL path
duration

Project duration = longest path; both must

Two paths critical Must shorten BOTH paths simultaneously
shrink

Any delay on this path = same delay to project

Critical path Always has Slack = 0
end

ALL paths must finish on time for project to

Project probability Multiply individual path probabilities
finish

Operations Research & Systems Analysis | Page 11

Project Management — All Math Types Explained URP 3291 | KUET

COMMON EXAM MISTAKES TO AVOID:

1. Taking SMALLEST EF on forward pass (wrong) -- always take LARGEST.
2. Taking LARGEST LS on backward pass (wrong) -- always take SMALLEST.
3. Forgetting to multiply ALL path probabilities -- not just critical path.
4. Crashing a non-critical activity and claiming it saves time.
5. Not checking whether a second path becomes critical after each crash.
6. Using z > 2.50 probability from table -- just write 1.000 directly.
7. Computing Slack as EF - ES (wrong) -- it must be LS - ES or LF - EF.

Operations Research & Systems Analysis | Page 12