Module 3: PERT Probabilistic Analysis

Module 3: PERT Probabilistic Analysis

CPM, Pert, and S-Curve Training Module

This module covers the Program Evaluation and Review Technique (PERT) applied to irrigation project management. It explains how to quantify schedule uncertainty using expected durations, variance, and standard deviation to determine the probability of meeting target completion dates.

Learning Objectives

Explain the difference between deterministic CPM and probabilistic PERT.
Define optimistic, most likely, and pessimistic time estimates in the context of typical irrigation construction uncertainties (e.g., weather, site access, equipment availability).
Calculate expected activity durations and variances.
Determine the total expected project duration and standard deviation along the critical path.
Calculate Z-scores and assess the probability of project completion by a target date to guide contingency planning and risk communication.

Program Evaluation and Review Technique (PERT)

A probabilistic project scheduling method designed to analyze and represent the tasks involved in completing a project, particularly when there is high uncertainty in individual activity durations. It uses three-point time estimates to calculate an expected duration and variance for schedule risk analysis.

Why PERT is Essential for Irrigation Infrastructure

Traditional Critical Path Method (CPM) relies on single, deterministic duration estimates. However, in reality, particularly in NIA-style irrigation projects, activity times vary due to unpredictable field constraints.

Common Uncertainty Drivers:

Weather Delays: Heavy rains halting earthworks or concrete pouring.
Site Access: Difficult terrain or lack of access roads during wet seasons.
Equipment/Logistics: Breakdown of heavy machinery, or delays in delivering concrete pipes to remote sites.
Geotechnical Issues: Unexpected hard rock encounters during main canal excavation.
Regulatory/Testing: Delays in securing right-of-way (ROW) permits, or waiting for material testing results (e.g., concrete cylinder compression tests).

PERT overlays a risk-based probabilistic model onto CPM, converting single guesses into statistically sound ranges, enabling better communication with stakeholders about realistic completion dates.

The Three-Point Estimate Framework

PERT captures duration uncertainty through three distinct time estimates for each activity:

Optimistic Time ( $a$ ): The minimum realistic time required if everything goes perfectly (excellent weather, highly productive crew, immediate material delivery).
Most Likely Time ( $m$ ): The normal, most realistic duration based on historical productivity rates and expected field conditions. This is often the value used in a standard CPM schedule.
Pessimistic Time ( $b$ ): The maximum realistic time required if significant but foreseeable delays occur (e.g., typical seasonal storms, minor equipment breakdowns, average material delays). Note: It should not include catastrophic, extremely rare events (like a 100-year flood or complete funding collapse), which are handled via separate risk registers, not schedule estimates.

Expected Activity Duration

PERT weighted average duration based on three-point estimates.

t_e = \frac{a + 4m + b}{6}

Variables

Symbol	Description	Unit
$t_e$	Expected duration of the activity.	-
$a$	Optimistic duration estimate (shortest realistic time).	-
$m$	Most likely duration estimate (normal condition time).	-
$b$	Pessimistic duration estimate (longest realistic time).	-

Understanding the Expected Duration Formula

The formula for expected duration is a weighted average based on the beta probability distribution. By giving the "most likely" time ( $m$ ) a weight of 4, and the extremes ( $a$ and $b$ ) a weight of 1 each, the formula heavily favors normal conditions while still slightly pulling the final average toward the pessimistic or optimistic end if the range is skewed. Because pessimistic estimates are typically further from the most likely estimate than optimistic ones (delays are more common than early completions), $t_e$ is often slightly longer than $m$ .

Adjust the three-point estimates in the interactive visualizer below to see how optimistic, most likely, and pessimistic values influence the expected duration and variance.

PERT Estimator Visualizer

Adjust the three-point estimates to see how they affect the expected duration and activity variance.

Optimistic Time

(a)

4 days

Most Likely Time

(m)

6 days

Pessimistic Time

(b)

10 days

Expected Duration

6.33 days

t_e = \frac{4 + 4(6) + 10}{6} = 6.33

Variance (

\sigma^2

)

1.00

Std Dev (

\sigma

)

1.00

Activity Variance

Measures the uncertainty or spread of the activity duration estimate.

\sigma^2 = \left(\frac{b - a}{6}\right)^2

Variables

Symbol	Description	Unit
$\sigma^2$	Variance of the activity duration.	-
$b$	Pessimistic duration estimate.	-
$a$	Optimistic duration estimate.	-

Significance of Activity Variance

Variance ( $\sigma^2$ ) quantifies the uncertainty of an activity. The larger the spread between the pessimistic and optimistic times, the higher the variance.

For example, routine clearing and grubbing might have a low variance because the time is highly predictable. Conversely, tunneling or deep excavation might have a high variance due to unknown sub-surface conditions. Project managers should prioritize close supervision and monitor daily progress reports heavily for high-variance activities that sit on the critical path.

Adjust the three-point estimates in the interactive visualizer below to see how optimistic, most likely, and pessimistic values influence the expected duration and variance.

Three-Point Estimator

Optimistic Duration (a): 4.0 days

Most Likely Duration (m): 6.0 days

Pessimistic Duration (b): 10.0 days

Calculated Results

Expected Duration (

t_e

)

t_e = \frac{4.0 + 4(6.0) + 10.0}{6} = 6.33 \text{ days}

Variance (

\sigma^2

)

\sigma^2 = \left(\frac{10.0 - 4.0}{6}\right)^2 = 1.00 \text{ days}^2

PERT Probabilistic Analysis Workflow

STEP-BY-STEP

Network Logic: Start with an established CPM activity list and network diagram.
Estimation: Gather optimistic, most likely, and pessimistic estimates for each activity from site engineers and historical project data.
Compute Activity Stats: Calculate the expected duration ( $t_e$ ) and variance ( $\sigma^2$ ) for every activity.
Determine Critical Path: Replace deterministic durations with the new expected durations ( $t_e$ ) and perform a forward and backward pass to identify the PERT critical path.
Project Variance: Identify the activities on the critical path and sum their individual variances to obtain the total project variance.
Project Standard Deviation: Take the square root of the total project variance to find the project standard deviation.
Probability Evaluation: Use the expected project duration, standard deviation, and standard normal distribution (Z-score) to calculate the probability of finishing by specific target dates (e.g., before the wet season).

The Critical Path Variance Rule

When calculating the total project variance, only sum the variances of the activities on the critical path.

Activities with float (non-critical activities) have buffer time to absorb delays. While high-variance non-critical activities still require monitoring—since significant delays could exhaust their float and create a new critical path—they do not directly dictate the overall statistical completion uncertainty of the project under normal circumstances.

Project Variance and Standard Deviation

Used to measure overall schedule uncertainty along the critical path.

V_p = \sum \sigma_{cp}^2

\sigma_p = \sqrt{V_p}

Variables

Symbol	Description	Unit
$V_p$	Total project variance.	-
$\sigma_{cp}^2$	Variance of an individual activity on the critical path.	-
$\sigma_p$	Project standard deviation.	-

Use the interactive simulation below to adjust the variance of individual critical path activities and observe how they accumulate into the total project variance and standard deviation.

Project Variance & Standard Deviation Aggregator

Adjust the individual variances of activities on the critical path to see how they aggregate into the total project uncertainty.

Activity 1 Variance (

\sigma_{1}^2

): 1.00 days²

Activity 2 Variance (

\sigma_{2}^2

): 0.44 days²

Activity 3 Variance (

\sigma_{3}^2

): 0.25 days²

Project Uncertainty

Project Variance (

V_p

)

V_p = 1.00 + 0.44 + 0.25 = 1.69 \text{ days}^2

Project Standard Deviation (

\sigma_p

)

\sigma_p = \sqrt{1.69} = 1.30 \text{ days}

Toggle specific risk events in the simulation below to evaluate their sensitivity and impact on schedule expected duration and variance.

Schedule Risk Sensitivity Analysis

Toggle specific risk events to see their impact on the overall expected duration and uncertainty of the project.

Adverse Weather During Earthworks
(+2.5 days

T_e

, +0.8 days² var)Delay in Concrete Pipe Delivery
(+1.0 days

T_e

, +0.3 days² var)Unexpected Hard Rock Excavation
(+3.5 days

T_e

, +1.5 days² var)

Risk-Adjusted Schedule

Expected Duration (

T_e

)

T_e = 38.0 + 0.0 = 38.0 \text{ days}

Total Variance (

V_p

)

V_p = 1.25 + 0.00 = 1.25 \text{ days}^2

Standard Deviation (

\sigma_p

)

\sigma_p = \sqrt{1.25} = 1.12 \text{ days}

Z-Score Probability and Risk Communication

Once you have the total expected duration and standard deviation, you can assess the likelihood of meeting external deadlines, such as funding expiration dates, VIP inaugurations, or seasonal weather shifts (e.g., completing dam embankment before monsoon season).

The Z-score measures how many standard deviations a target date is away from the expected completion date. This relies on the Central Limit Theorem, which assumes that the sum of independent activity durations along the critical path approximates a normal (bell-shaped) distribution curve.

Completion Probability Z-Score

Used to evaluate the probability of completing the project by a target date using standard normal distribution.

Z = \frac{T_d - T_e}{\sigma_p}

Variables

Symbol	Description	Unit
$Z$	Z-score representing the number of standard deviations the target date is from the expected date.	-
$T_d$	Target project duration or completion date dictated by management or external factors.	-
$T_e$	Total expected project duration (sum of expected durations along the critical path).	-
$\sigma_p$	Project standard deviation.	-

Explore the interactive visualizer below to see how expected duration, target date, and standard deviation affect the probability of completing the project on time.

Completion Probability Analysis

Expected Duration (

T_e

): 40.0 days

Target Duration (

T_d

): 43.0 days

Standard Deviation (

\sigma_p

): 2.0 days

Z-Score

Z = \frac{43.0 - 40.0}{2.0} = 1.50

Probability

93.32%

Use the interactive planner below to set your required confidence level and determine the necessary schedule contingency for your target completion date.

Contingency Decision Planner

Set your required confidence level to determine the schedule contingency needed to commit to a target date.

Desired Confidence Level: 85%

(Baseline $T_e = 40$ , $\sigma_p = 2$ )

Required Target Date

Required Z-Score

Z = 1.04

Calculated Contingency

\text{Contingency} = Z \times \sigma_p = 1.04 \times 2.0 = 2.1 \text{ days}

Committed Target Date (

T_d

)

T_d = T_e + \text{Contingency} = 40.0 + 2.1 = 42.1 \text{ days}

Connecting PERT to Management Decision Thresholds

PERT outputs are not just theoretical numbers; they directly inform project management action:

High Variance: Triggers early contingency planning. If pipe delivery variance is high, management might authorize purchasing from two suppliers simultaneously.
Low Probability of Success (e.g., < 50%): If the Z-score analysis shows a low probability of hitting the target, the project manager must communicate this risk proactively to stakeholders. It justifies requesting additional budget for schedule compression (crashing by adding crews/shifts or fast-tracking by running tasks in parallel).
High Probability of Success (e.g., > 95%): Indicates the target date is highly achievable, which might allow the project team to release reserve resources or equipment to other delayed phases of the project.

PERT Data Quality Checklist

Three-point estimates are derived from historical data, standard productivity manuals, or consensus among experienced site engineers, not arbitrary guesses.
Specific uncertainty drivers (weather, ROW, delivery) are explicitly documented alongside the optimistic and pessimistic estimates.
Pessimistic estimates include foreseeable risks but strictly exclude catastrophic "Acts of God" (force majeure).
The network's critical path is recalculated strictly using $t_e$ , not $m$ .
Z-score probabilities are translated into plain-language risk reports and actionable recovery plans for management review.

Explore the comprehensive PERT Probabilistic Diagram Visualizer below to see how node estimates build into expected durations, variances, and ultimately identify the critical path uncertainties.

PERT Probabilistic Diagram Visualizer

Critical Path

Non-Critical

Node B: Excavation

Optimistic (a)4.0

Most Likely (m)6.0

Pessimistic (b)10.0

Project Totals (Critical Path)

Total Exp. Duration:20.00 d

Total Variance:2.72 d²

Project Std Dev:1.65 d

Key Takeaways

PERT transforms deterministic single-point CPM schedules into probabilistic risk-management tools.
Expected duration ( $t_e$ ) is a weighted average ( $\frac{a+4m+b}{6}$ ), heavily influenced by the most likely time but pulled slightly by extreme estimates.
Activity variance ( $\sigma^2$ ) highlights which tasks carry the highest uncertainty and require the strictest site supervision.
Total project risk is evaluated by summing the variances of only the critical path activities.
The Z-score maps the expected duration and standard deviation to the probability of meeting hard target dates, directly informing contingency funding and schedule compression decisions.

Previous TopicCPM Fundamentals - Worked Examples

Quiz Me

Next TopicPERT Probabilistic Analysis - Worked Examples

Prev Next

Quiz Me