Critical Path Method and Network Analysis

Introduction and context

Construction projects consist of interdependent activities performed by different trades, suppliers, consultants, and authorities. A schedule is more than a list of dates: it is a logical model of how work must proceed. Network analysis represents this logic and identifies the activities that control project completion.

The Critical Path Method (CPM) is a deterministic scheduling technique. It uses known or assumed activity durations, precedence relationships, and a project completion objective to calculate the earliest and latest permissible dates for each activity. It also identifies float and critical paths. CPM is particularly useful during preconstruction planning, procurement coordination, progress monitoring, delay analysis, and recovery planning.

A CPM network does not automatically create a realistic schedule. Its quality depends on sound activity definition, reliable durations, correct logic, appropriate calendars, and disciplined updating.

Learning objectives

By the end of this chapter, a student should be able to:

1. Construct an activity-on-node (AON) network.
2. Explain finish-to-start, start-to-start, finish-to-finish, and start-to-finish relationships.
3. Perform forward and backward passes.
4. Calculate early dates, late dates, total float, and free float.
5. Identify critical paths and explain multiple-path sensitivity.
6. Apply constraints, calendars, lags, and leads appropriately.
7. Interpret CPM outputs for construction control and decision-making.
8. Recognize the assumptions and limitations of deterministic network analysis.

1. Network planning concepts

Activities, events, and logic

An activity is a measurable item of work requiring time and usually resources, such as “install reinforcement” or “cure concrete.” In AON notation, each activity is represented by a node or box. Arrows represent logical relationships between activities.

A predecessor is an activity that must satisfy a relationship before another activity can proceed. A successor is the activity that follows it. A network should have a clear start and finish, although a practical schedule may contain several activities beginning at the data date or several completion milestones.

A network is a directed graph. It should normally be acyclic: following the arrows must not lead back to an earlier activity. A circular relationship, such as A depending on B while B depends on A, is a scheduling error.

Activity-on-node notation

A simple node may contain the following fields:

In a basic finish-to-start network, the activity duration is measured in consistent units such as working days. If an activity starts at time 0 and lasts 5 working days, its finish is time 5. If a schedule uses day numbers beginning at Day 1, the display convention may differ by one day; the underlying time intervals must remain consistent.

Figure: A simplified activity-on-node network; red nodes illustrate a controlling path.

Precedence relationships

The four standard relationships are:

FS is the most frequently used relationship. A lag is a deliberate waiting or separation period, such as a two-day curing interval. A lead permits overlap and should be used cautiously because it can conceal required work or create unsafe assumptions.

Network rules

A reliable network should:

• include all activities necessary to define the chosen scope;
• use logically defensible relationships rather than arbitrary dates;
• avoid excessive constraints and unexplained lags;
• identify procurement, approvals, inspections, testing, and handover activities;
• contain no unintended loops;
• distinguish work activities from milestones and imposed dates; and
• be traceable to the work breakdown structure and schedule basis.

2. Forward pass: calculating early dates

The forward pass determines the earliest dates at which activities can start and finish, assuming predecessors occur as early as permitted.

For a basic FS relationship with zero lag:

where ES_i is early start of activity i, EF_i is early finish, and D_i is duration. All three quantities must use the same time unit.

For an activity with several predecessors in a pure FS network:

where the maximum is taken over all predecessors i of activity j, and L_ij is the lag. For a project start activity, ES = 0 is commonly used. The formula assumes FS logic and a consistent working calendar.

For other relationship types, the controlling predecessor expression changes. In simplified time coordinates:

• FS: successor start ≥ predecessor finish + lag.
• SS: successor start ≥ predecessor start + lag.
• FF: successor finish ≥ predecessor finish + lag.
• SF: successor finish ≥ predecessor start + lag.

The scheduler should then derive the successor’s other date from its duration. Commercial scheduling software performs these operations using calendars and date conventions; manual calculations are safest for simple, common-calendar examples.

3. Backward pass: calculating late dates

The backward pass calculates the latest dates that activities may occur without delaying the required project completion date.

For a basic FS network:

where LS_i is late start and LF_i is late finish.

For an activity with several successors:

where the minimum is taken over all successors j. The final activity’s late finish is set equal to the required project completion time, often the calculated earliest project finish when no external deadline exists.

These formulas assume a simple network, one consistent calendar, FS logic, and no special constraints. When calendars differ, dates should be calculated in working-time units using the applicable calendar rather than by ordinary calendar subtraction.

4. Float and criticality

Total float

Total float is the amount of time an activity can be delayed without delaying the schedule completion date, assuming the rest of the network and the required completion date remain unchanged.

Total float is expressed in working days under the relevant schedule calendar. Small rounding or calendar differences may produce apparently unequal forms in software; this usually indicates differing calendars or date conventions and should be investigated.

Activities with approximately zero total float form a critical path. In practice, “critical” should not be treated as a permanent label. It is a status resulting from the current logic, durations, calendars, constraints, and completion target.

Free float

Free float is the amount of time an activity can be delayed without delaying the earliest start of any immediate successor.

For a basic FS network:

where j represents immediate successors. For a terminal activity, free float is often taken as the difference between the project’s required completion and its early finish, although organizational conventions vary.

Free float cannot exceed total float in a correctly constructed basic network. Total float concerns project completion; free float concerns immediate successors.

5. Worked example: a small building package

Consider the following illustrative activities, all measured in working days and using FS relationships with zero lag.

Forward pass

• A: ES 0, EF 3.
• B: ES 3, EF 8.
• C: ES 3, EF 10.
• D: ES max(8, 10) = 10, EF 14.
• E: ES 14, EF 20.
• F: ES 8, EF 11.
• G: ES max(11, 14) = 14, EF 16.
• H: ES max(20, 16) = 20, EF 20.

The earliest project completion is 20 working days.

Backward pass

Set H LF = 20 and LS = 20.

• E: LF 20, LS 14.
• G: LF 20, LS 18.
• D: LF min(14, 18) = 14, LS 10.
• F: LF 18, LS 15.
• B: LF min(10, 15) = 10, LS 5.
• C: LF 10, LS 3.
• A: LF min(5, 3) = 3, LS 0.

Float calculation

For example, B has TF = 5 - 3 = 2 days. Its free float is min(ES_D, ES_F) - EF_B = min(10, 8) - 8 = 0. Activity F has total float 7 days and free float ES_G - EF_F = 14 - 11 = 3 days.

The critical path is A–C–D–E–H, with duration 20 working days. Although B is not critical, its two-day total float is modest. A delay of three days to B would consume all its float and create a new controlling path through B and D. This illustrates why noncritical activities still require management attention.

6. Multiple critical paths and sensitivity

A project may have two or more paths with equal or nearly equal durations. If paths are:

• Path 1: 30 days;
• Path 2: 29 days; and
• Path 3: 22 days,

Path 1 is critical, while Path 2 has only one day of path float. The schedule is highly sensitive because a small delay on Path 2 can make it critical.

Multiple-path sensitivity is the degree to which alternative paths are close to the controlling path. A useful practical measure is the difference between the longest path and another significant path:

where T_max is the longest path duration and T_k is the duration of path k. This is not a substitute for activity-level float, but it helps identify fragile schedules.

Management responses include monitoring near-critical activities, protecting procurement and approvals, developing recovery options, and testing scenarios. A schedule with one apparently dominant path may still be risky if calendars, productivity, or logic are uncertain.

7. Constraints, calendars, and lags

Constraints

A constraint limits the dates available to an activity. Examples include a contractual completion date, a possession date, a permit release, or a client-imposed access window.

Constraints should reflect genuine project conditions. A strong or mandatory date constraint can override network logic and create negative float if the required date is earlier than the calculated completion. Negative float does not mean that work can be performed in negative time; it indicates a shortfall against the imposed requirement.

Prefer logic-based relationships where the condition is caused by another activity. For example, link “permit approved” to “excavation starts” rather than assigning an unexplained start-no-earlier-than date. Constraints are sometimes necessary, but excessive constraints reduce transparency and can hide the true driving path.

Calendars

A calendar defines working days, nonworking days, shifts, holidays, working hours, and sometimes resource-specific availability. Common calendars include project, activity, and resource calendars.

If a five-day activity begins on a Friday under a Monday–Friday calendar, the finish is not obtained by adding five ordinary calendar days. The calculation must count the applicable working intervals. Distinct calendars can produce different early and late dates even when nominal durations are identical.

The schedule basis should document holidays, planned shutdowns, weather allowances, shift patterns, time zones where relevant, and the calendar assigned to each activity. Calendar changes during updating can create apparent progress or delay and must be controlled.

Lags and leads

A lag represents time between related activities, such as curing, statutory review, or shipping. Lags should have a documented basis and preferably be represented as explicit activities when they require monitoring or resources.

A lead creates overlap. For example, wall framing may begin before all slab finishing is complete if the work fronts are physically separated. Leads can be difficult to explain and audit; explicit overlapping activities or realistic SS/FF relationships are usually clearer.

8. Construction application and control

CPM is used during:

• baseline schedule development;
• subcontractor and procurement integration;
• look-ahead planning;
• extension-of-time and delay analysis;
• recovery and acceleration studies;
• coordination of inspections and approvals; and
• forecasting of completion dates.

A schedule update compares actual starts, actual finishes, remaining durations, and revised logic with the baseline. The updated critical path may differ from the baseline path. A delay to a noncritical activity may become critical after float is consumed; conversely, acceleration may move control to another path.

For delay analysis, the analyst must preserve contemporaneous records: approved schedules, updates, site diaries, instructions, submittal registers, productivity data, weather records, and correspondence. A critical-path conclusion should be supported by the schedule status at the time of the event, not merely by a retrospective manipulation.

9. Illustrative case: municipal library construction

Background

A contractor is constructing a municipal library with a 42-working-day structural and envelope phase. The baseline has one critical path through foundations, frame, roof, and enclosure. Mechanical equipment procurement and electrical containment form a second path ending two days later than the critical path’s successor sequence would permit.

Problem

A design clarification delays equipment release by four working days. At the same time, unusually wet weather reduces excavation productivity. The project team initially treats equipment procurement as noncritical because it had two days of float.

Analysis

The procurement delay consumes the two days of float and creates a two-day delay to the project unless the equipment path is recovered. The weather delay consumes the small float on excavation and threatens the foundation path. The schedule now has two near-critical paths, so focusing only on the original red path would be misleading.

Approach and solution

The planning team validates the logic, confirms the approval dates, and separates procurement release, fabrication, delivery, and installation rather than using one long activity. It evaluates three options:

1. expedite fabrication at additional cost;
2. resequence containment and nondependent internal work; and
3. add a second excavation crew where site access permits.

The team selects limited fabrication expediting and resequencing. It records the approved change, updates remaining durations, and reports both path risks at the weekly coordination meeting.

Outcome

The equipment path loses its float but does not delay final completion. Excavation recovery restores one day of float on the foundation path. The project finishes one day later than the original baseline but within the revised contractual allowance.

Learning points

The case demonstrates that float is a management resource, not unused time; near-critical paths require active monitoring; procurement logic must be detailed enough to expose release and delivery risk; and schedule decisions should consider cost, safety, quality, and contractual entitlement together.

10. Limitations, governance, and technology

CPM is deterministic. It normally uses one duration for an activity, even though actual construction durations are uncertain. It also assumes that logical relationships adequately represent work and that resources are available unless resource constraints are explicitly modelled. A mathematically correct network can therefore be operationally impossible.

Resource leveling may delay activities and alter the critical path. In such cases, the resource-constrained schedule should be distinguished from the purely logic-driven CPM result. Weather, productivity, design change, supply disruption, and safety events require scenario analysis rather than blind reliance on baseline dates.

Professional governance requires version control, an approved schedule basis, documented assumptions, traceable updates, and review by people who understand the work. Artificially subdividing activities, inserting unnecessary constraints, or manipulating durations to produce a desired completion date undermines professional integrity and may create contractual and legal risk.

Scheduling software can automate date calculations, calendars, baselines, resource loading, and what-if analysis. Artificial intelligence can assist with activity coding, anomaly detection, progress extraction from site images, and delay-risk forecasting. However, AI-generated logic or durations must be checked against drawings, specifications, method statements, site conditions, safety requirements, and contractual obligations. Automation does not transfer accountability from the competent planner or project manager.

Best-practice checklist

Before approving a CPM network, verify that:

1. The scope is aligned with the work breakdown structure.
2. Activities have measurable outputs and credible durations.
3. Logic reflects actual construction methods and interfaces.
4. Procurement, approvals, inspections, testing, and handover are represented.
5. Calendars and holidays are documented.
6. Constraints and lags have written justification.
7. The network has no unintended loops or open ends.
8. Forward and backward pass results have been reviewed.
9. Critical and near-critical paths are reported.
10. Updates use actual records and preserve the baseline.
11. Resource availability and safety constraints are considered.
12. The schedule is communicated in a form usable by supervisors and trades.

Chapter Summary

Critical Path Method models construction work as a network of activities and logical relationships. In an activity-on-node network, the forward pass calculates early dates and the backward pass calculates late dates. Total float measures delay permitted without delaying required project completion, while free float measures delay permitted without affecting the early start of an immediate successor. Activities with zero or near-zero total float form critical or near-critical paths.

Correct CPM practice requires more than arithmetic. The planner must select suitable activities, represent construction logic, apply calendars, document constraints and lags, and interpret multiple-path sensitivity. Criticality can change as work progresses, resources become constrained, or actual durations differ from assumptions. CPM is therefore both a calculation method and a management discipline. Its outputs should support transparent decisions about procurement, sequencing, recovery, risk, and contractual control.

Key Terms

Activity: A defined item of work requiring time and usually resources.

Activity-on-node: A network convention in which activities are shown as nodes and relationships as arrows.

Backward pass: Calculation of latest allowable activity dates.

Calendar: Definition of working and nonworking time.

Constraint: A restriction imposed on an activity’s timing.

Critical path: Longest controlling path through a network under stated conditions.

Critical activity: Activity with approximately zero total float.

Early finish: Earliest permissible completion time of an activity.

Early start: Earliest permissible start time of an activity.

Float: Schedule flexibility available to an activity or path.

Free float: Delay allowed without delaying an immediate successor’s early start.

Forward pass: Calculation of earliest activity dates.

Lag: Time separation imposed between related activities.

Lead: Permitted overlap between related activities.

Late finish: Latest completion time that preserves the required project completion.

Late start: Latest start time that preserves the required project completion.

Near-critical path: A path with little float relative to the controlling path.

Predecessor: Activity that logically precedes another activity.

Successor: Activity that logically follows another activity.

Total float: Delay allowed without delaying required project completion.

Review Questions

Short Answer Questions

1. What is the purpose of a forward pass in CPM?
2. Distinguish total float from free float.
3. Define a finish-to-start relationship and give one construction example.
4. Why should unexplained constraints be avoided?
5. What does negative float indicate?

Long Answer Questions

1. Explain how an AON network is developed from a construction work breakdown structure.
2. Describe the forward-pass and backward-pass procedures, including the conditions under which the standard formulas apply.
3. Explain how calendars, lags, leads, and constraints affect CPM results.
4. Discuss why a noncritical activity may still require close management.
5. Evaluate the main limitations of deterministic CPM in construction projects.

Analytical / Application Questions

1. Construct an AON network for a small concrete-frame project containing excavation, reinforcement procurement, foundations, columns, beams, curing, and inspection. Identify all logical relationships and justify any lags.
2. Using the activity data in the worked example, determine which activity would become critical if activity B increased from 5 to 8 working days. Explain the effect on project completion.
3. A project has paths of 36, 35, and 27 working days. Explain the sensitivity of the schedule and propose three management actions for controlling completion risk.

Multiple Choice Questions with Answers

Question 1

In a basic finish-to-start network, which expression calculates an activity’s early finish?

A. EF = ES + duration

B. EF = LS − duration

C. EF = LF + duration

D. EF = ES − duration

Correct Answer: A. EF = ES + duration

Question 2

For an activity with several finish-to-start predecessors, its early start is normally based on which value?

A. The minimum predecessor early finish

B. The average predecessor early finish

C. The maximum predecessor early finish plus applicable lag

D. The latest predecessor late start

Correct Answer: C. The maximum predecessor early finish plus applicable lag

Question 3

Which statement best defines total float?

A. Time available before an activity begins

B. Delay permitted without delaying required project completion

C. Time between two unrelated activities

D. Duration of the longest activity

Correct Answer: B. Delay permitted without delaying required project completion

Question 4

Using TF = LS − ES, an activity with ES = 6 and LS = 10 has total float of:

A. 4 working days

B. 6 working days

C. 10 working days

D. 16 working days

Correct Answer: A. 4 working days

Question 5

Which relationship means that a successor may start only after its predecessor finishes?

A. Start-to-start

B. Finish-to-finish

C. Start-to-finish

D. Finish-to-start

Correct Answer: D. Finish-to-start

Question 6

In a correctly calculated basic CPM network, an activity with zero total float is generally described as:

A. Independent

B. Critical

C. Optional

D. Overallocated

Correct Answer: B. Critical

Question 7

If an activity has EF = 12 and the earliest start of its immediate successor is 15 under a basic FS relationship, its free float is:

A. 3 working days

B. 12 working days

C. 15 working days

D. 27 working days

Correct Answer: A. 3 working days

Question 8

What does negative float usually indicate?

A. The activity has no predecessor

B. The imposed completion requirement is earlier than the calculated schedule can achieve

C. The activity duration is zero

D. The project has excessive available time

Correct Answer: B. The imposed completion requirement is earlier than the calculated schedule can achieve

Question 9

A project has paths of 40 and 39 working days. What is the principal scheduling implication?

A. The 39-day path can be ignored

B. Both paths deserve close monitoring because the second is near critical

C. The project must finish in 79 days

D. Neither path can become critical

Correct Answer: B. Both paths deserve close monitoring because the second is near critical

Question 10

Which practice most improves the transparency of a construction schedule?

A. Using unexplained mandatory dates for most activities

B. Removing procurement and approval activities

C. Documenting calendars, assumptions, logic, constraints, and update records

D. Changing durations to match a desired completion date

Correct Answer: C. Documenting calendars, assumptions, logic, constraints, and update records