Schedule and Network Planning Foundations

Introduction

Resource loading and leveling are only as reliable as the schedule on which they are based. A resource histogram can show apparent over-allocation, but it cannot determine whether the underlying work sequence is logically valid. Similarly, a leveling algorithm may delay activities efficiently while concealing an incorrect dependency, an unrealistic calendar, or an omitted handover. Schedule and network planning therefore constitute the analytical foundation for resource analysis.

This chapter develops the concepts required to construct, analyse, validate, and govern project schedules before resources are loaded. It focuses on precedence networks, the Critical Path Method (CPM), dependency types, leads and lags, calendars, float, critical paths, milestones, baselines, and schedule quality. The emphasis is both computational and managerial: a network is not merely a diagram but a formal representation of how project work is expected to unfold.

Learning Objectives

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

1. Translate a work breakdown structure and activity information into a precedence network.
2. Distinguish activity-on-node notation and the four principal dependency types.
3. Perform forward-pass and backward-pass calculations using CPM.
4. Interpret total float, free float, critical paths, near-critical paths, and milestones.
5. Apply calendars, leads, lags, and constraints without corrupting schedule logic.
6. Establish and use a schedule baseline.
7. Diagnose schedule-quality problems before resource loading and leveling.
8. Evaluate the managerial implications and limitations of network-based scheduling.

1. The Schedule as a Project-Control Model

A project schedule is a time-based model of scope, sequence, duration, constraints, and, ultimately, responsibility. It answers at least five questions:

• What work must be performed?
• In what logical order can or must it occur?
• How long is each activity expected to take?
• Which dates and milestones matter to stakeholders?
• How much flexibility exists before completion or another commitment is affected?

The schedule should be distinguished from the work breakdown structure (WBS). The WBS organizes deliverables and scope; the schedule organizes the activities needed to produce those deliverables over time. A WBS element may contain several schedule activities, and a schedule activity may contribute to more than one deliverable.

A credible schedule normally contains activity identifiers, descriptions, estimated durations, relationships, calendars, constraints, responsible parties, milestones, and status information. For resource analysis, it must also support assignment of labor, equipment, materials, facilities, and cost. If the network is incomplete, resource analysis will be incomplete even when the software output appears precise.

Important terminology

An activity is a defined piece of work with a measurable duration. A predecessor is an activity that logically precedes another activity; a successor follows it. A node is a graphical representation of an activity or event. An arc or arrow represents a logical relationship. A path is a sequence of connected activities through the network. A milestone is a zero-duration event or significant point used for control, approval, or communication.

2. From Scope to a Precedence Network

Network development begins with scope decomposition. The scheduler identifies work packages, decomposes them into schedule activities, estimates durations, and then asks what must happen before, after, or concurrently with each activity. Relationships should represent necessary logic rather than preferred habits or resource convenience.

The most common modern notation is activity-on-node (AON), also called precedence diagramming. Each activity appears in a box or node, while arrows show dependencies. AON supports four relationship types and is well suited to contemporary scheduling software. Older activity-on-arrow systems represent activities as arrows and events as nodes; they may require dummy activities to express certain relationships. Unless otherwise stated, CPM calculations in this chapter use AON.

A basic network should have a clear start and finish, although large programs may have multiple contractual or control milestones. Activities should generally be connected to the network. Excessive use of date constraints, unexplained open ends, and unnecessary summary-task relationships are warning signs.

Figure: An activity-on-node network in which installation can begin only after both procurement and site preparation are complete.

3. Dependency Types and Network Logic

The four standard logical relationships are as follows.

Finish-to-start is the default and most frequently used relationship. The other types are legitimate when they represent actual work logic, but they should not be used merely to force a desired date.

A lead allows a successor relationship to overlap its predecessor. A lag inserts waiting time or separation. For an FS relationship with lag L, the successor start cannot occur before the predecessor finish plus the lag. A positive lag might represent curing, statutory review, shipping, or stabilization. A lead is often represented as a negative lag, although organizations may display it separately to improve governance.

Leads and lags require careful interpretation. A statement such as “start installation two days before design is complete” may conceal unfinished inputs and should be tested with technical staff. Excessive lags are particularly problematic because they hide work and make the schedule difficult to update. Where waiting is a real activity, it may be better to model it explicitly.

For a relationship from activity i to activity j, a generalized precedence condition can be written as:

Here, S_i and S_j are start times, d_i is the duration of predecessor i, and L_ij is the lag; time units must be consistent. This expression applies directly to an FS relationship. For SS, FF, and SF relationships, the relevant start or finish terms must be used according to the relationship definition. The equation expresses a logical minimum, not necessarily the actual scheduled date.

4. Calendars and Time Representation

A calendar determines which periods are available for work. Common calendars include a five-day workweek, a seven-day calendar, a shift calendar, a plant shutdown calendar, and an individual resource calendar. Calendars may include working hours, weekends, public holidays, maintenance windows, and local time-zone rules.

Duration and elapsed time are not always equivalent. A five-working-day activity beginning on Monday may finish on Friday under a standard calendar, whereas five elapsed days may include a weekend. A calendar can therefore change both dates and apparent float without changing the nominal duration.

The scheduling team should distinguish:

• Activity calendar: when the activity is permitted to occur.
• Resource calendar: when a particular resource is available.
• Project calendar: the default working pattern for the project.
• Exception: a holiday, shutdown, overtime period, or special working day.

Calendar assumptions must be documented. Applying a resource calendar to an activity before resources have been assigned can create circular reasoning: the activity duration appears to depend on resource availability, while resource availability is being analysed from the activity dates. In resource loading, this distinction becomes especially important.

5. Critical Path Method

CPM is a deterministic network-analysis technique. It uses specified activity durations and logical relationships to calculate earliest and latest feasible dates. It identifies paths that govern the calculated project completion date under the stated assumptions.

CPM does not prove that the duration estimate is correct. It also does not automatically account for uncertain durations, resource contention, weather, or changing scope. Its results are conditional on the network, durations, calendars, and constraints supplied.

5.1 Forward pass

For a simple FS network using a time origin of zero, the earliest start and earliest finish are calculated as follows:

ES_i is the earliest start of activity i, EF_i its earliest finish, d_i its duration, and L_ij the lag from predecessor i to successor j. The maximum is taken over all immediate predecessors of j. The equations assume a consistent time unit and an FS relationship. For a start-at-one convention, the organization must adjust the interpretation of the time origin consistently.

5.2 Backward pass

The backward pass calculates the latest dates that preserve the required project completion date. For a project completion time T, the final activity’s latest finish is set to T. Then:

LS_i and LF_i are latest start and latest finish; the minimum is taken over all immediate successors. Again, this form assumes FS relationships and compatible lag treatment. Other dependency types require the corresponding transformed relationship equations.

5.3 Float

Total float measures how much an activity can move without delaying the calculated project completion date, assuming the remainder of the network remains as planned:

Total float is measured in time units. In a standard unconstrained CPM network, activities on a critical path have zero total float, subject to rounding and calendar conventions. In practice, constraints, multiple completion milestones, and software settings can produce negative or nonzero float on activities that managers still treat as critical.

Free float is the amount an activity can be delayed without delaying the earliest start of any immediate successor. For a simple FS network:

where j ranges over immediate successors. Free float cannot exceed total float in the usual forward/backward-pass interpretation. Float is not a reserve of labor or budget; it is a property of schedule logic under stated assumptions.

6. Worked CPM Example

Consider the following illustrative activities. Durations are working days, and all relationships are FS with zero lag.

Using a time origin of zero, the forward pass is:

• A: ES = 0; EF = 0 + 4 = 4.
• B: ES = 4; EF = 4 + 3 = 7.
• C: ES = 4; EF = 4 + 5 = 9.
• D: ES = max(7, 9) = 9; EF = 9 + 6 = 15.
• E: ES = 15; EF = 15 + 4 = 19.
• F: ES = 7; EF = 7 + 3 = 10.
• G: ES = max(19, 10) = 19; EF = 19 + 2 = 21.

The calculated project duration is 21 working days. The backward pass begins with G having LF = 21 and LS = 19:

• G: LF = 21; LS = 21 − 2 = 19.
• E: LF = 19; LS = 19 − 4 = 15.
• F: LF = 19; LS = 19 − 3 = 16.
• D: LF = 15; LS = 15 − 6 = 9.
• C: LF = 9; LS = 9 − 5 = 4.
• B: LF = min(9, 16) = 9; LS = 9 − 3 = 6.
• A: LF = min(6, 4) = 4; LS = 4 − 4 = 0.

The resulting values are:

The critical path is A–C–D–E–G, with a duration of 21 working days. Activity B has two days of total float, while F has nine days. Notice that B is not critical even though it is a predecessor of D; C controls D because C finishes later than B. This is an important managerial insight: the longest path through the network, not the path with the greatest number of activities, governs completion.

If an additional contractual milestone were inserted after training, F might become critical for that milestone even though it is not critical for final deployment. Criticality is therefore relative to a defined completion event.

7. Criticality, Near-Criticality, and Schedule Risk

The critical path is often treated as a single immutable chain, but real schedules may contain multiple critical paths, paths that change as progress occurs, and near-critical paths with little float. A two-day path can become critical after a minor delay. Near-critical paths deserve management attention because they may have higher uncertainty, scarce resources, or strong dependencies on external parties.

A deterministic critical path is not identical to a statistically most likely path. In probabilistic scheduling, activity durations may be represented by distributions and analysed through simulation. The probability that a path becomes critical depends on duration uncertainty, correlation, calendars, and network structure. PERT-style three-point estimates can be useful as a communication device, but they should not be treated as a substitute for empirical estimating or simulation when risk is material.

Resource constraints can also change the controlling sequence. A purely logic-based CPM network may show parallel activities, but if both require the same scarce specialist, they cannot actually proceed simultaneously. Resource leveling introduces additional temporal restrictions and may create a resource-driven critical path. This is why logic-based CPM should precede, and then be revisited after, resource analysis.

8. Milestones, Constraints, and Baselines

A milestone has zero duration but significant control meaning. Examples include design approval, contract award, equipment arrival, regulatory acceptance, readiness for commissioning, and client handover. Milestones should be connected to the activities that produce them. A milestone floating independently with a hard date provides little explanatory value.

Constraints are imposed date conditions. Examples include a required finish date, a no-earlier-than date, a no-later-than date, or a contractual deadline. Constraints may be necessary, but excessive hard constraints can mask network logic and create misleading float. The preferred practice is to model the causal logic and use constraints only where a real external condition exists.

A baseline is an approved version of the schedule against which actual performance and forecasts are compared. A baseline normally preserves planned dates, durations, logic, milestones, and sometimes planned resource and cost values. It is not a frozen plan that can never change; it is a controlled reference. Approved changes should be incorporated through documented change control, with the original baseline retained for historical analysis where governance requires it.

Baseline comparisons should distinguish:

• variance caused by actual progress;
• variance caused by approved scope or sequence changes;
• variance caused by estimating error;
• variance caused by calendar or status-date changes;
• variance caused by resource or external constraints.

Changing the baseline to eliminate unfavorable variance undermines governance and prevents organizational learning.

9. Schedule Quality Checks Before Resource Loading

A schedule-quality review should occur before assigning detailed resources. Typical checks include the following.

Logical completeness

Every non-start activity should have a predecessor, and every non-finish activity should have a successor, unless an exception is documented. Open ends can make dates appear flexible when they are not supported by logic.

Relationship quality

Review excessive SS, FF, SF, leads, and lags. A relationship should have a clear technical or contractual explanation. Summary tasks should not normally be used as substitutes for detailed activity logic.

Duration realism

Extremely long activities reduce visibility and make progress measurement difficult. Extremely short activities may create administrative noise. Durations should be consistent with the estimating method, work content, production assumptions, and calendar.

Constraint and float review

A large number of hard constraints, negative float values, or unexpectedly high float values may indicate incorrect logic. Negative float usually means that the calculated schedule cannot meet an imposed date under current assumptions; it is a signal for management action, not a scheduling error to be hidden.

Critical-path review

The critical path should be technically credible and should lead to a meaningful completion milestone. Review near-critical paths, convergence points, and activities with high uncertainty.

Calendar consistency

Check holidays, shutdowns, shifts, time zones, daylight-saving changes where relevant, and resource-specific availability. A calendar error can create systematic date errors throughout the network.

Status and progress integrity

Actual starts and finishes should be based on evidence. Remaining duration should not be replaced automatically by percentage complete. Physical progress, earned quantities, inspection records, and accepted deliverables are often better evidence than subjective percentages.

Traceability and auditability

Activity identifiers, assumptions, approvals, changes, and data dates should be traceable. A schedule that cannot explain why a date exists is difficult to defend in governance forums or disputes.

Schedule-quality metrics can be formalized by an organization, but no single score proves that a schedule is good. A network may satisfy software checks and still be operationally unrealistic. Expert review remains essential.

10. Applied Case: Data-Centre Commissioning Schedule

Background

The following is a fictional illustrative case. A project team is preparing a small data-centre facility for client handover. The resource manager wants to load electricians, commissioning engineers, safety inspectors, and client representatives into a scheduling tool.

Problem

The initial schedule shows commissioning completing in 18 working days. However, it contains a single summary activity called “Install systems,” a 12-day lag before testing, and a finish-date constraint on handover. The resource histogram suggests that commissioning engineers are underutilized, while the project manager believes the handover date is at risk.

Analysis

The planning team decomposes the summary activity into electrical installation, cooling installation, controls configuration, integrated testing, defect correction, client demonstration, and handover. Interviews reveal that controls configuration can begin after controls panels are installed, but integrated testing requires both electrical and cooling systems to be accepted. The 12-day lag is discovered to represent several distinct conditions: two days for insulation stabilization, five days for vendor availability, three days for safety inspection, and two days of undocumented contingency.

The team models the first three conditions explicitly. Vendor availability is represented as a real external milestone, while the undocumented contingency is removed from the logic and recorded as a risk response allowance rather than hidden lag. A seven-day working calendar is applied to specialist commissioning activities because vendor staff work weekends, but client acceptance remains on a five-day calendar.

The revised network identifies two near-critical paths: cooling installation–stabilization–integrated testing and controls configuration–functional testing–integrated testing. Both converge at integrated testing. The resource analysis then reveals that one commissioning engineer is required on both functional and integrated testing. The team decides to sequence selected test packages and add a second engineer for three days.

Approach and outcome

After logic correction, the calculated completion date moves from day 18 to day 22, exposing the previously hidden risk. Adding the second engineer and obtaining an earlier vendor slot reduces the forecast to day 20. The revised schedule is then baselined after stakeholder approval. Resources are loaded only after the activity logic and calendars are accepted.

Learning points

1. A summary activity concealed the actual dependency structure.
2. A long lag mixed technical waiting, external availability, inspection, and contingency.
3. Multiple near-critical paths converged at a testing milestone.
4. Resource loading revealed a genuine capacity constraint only after the network was corrected.
5. A later but credible baseline is more valuable than an earlier schedule that hides risk.

11. Technology, Automation, and Governance

Scheduling software can calculate dates, display logic, perform network analyses, compare baselines, and generate resource histograms. Automated schedule checks can identify open ends, excessive constraints, invalid relationships, out-of-sequence progress, and unusually high float. These functions improve consistency but do not replace professional judgment.

AI-assisted scheduling is increasingly used to suggest activity decompositions, infer relationships from historical projects, detect anomalous durations, and propose leveling alternatives. Such systems should be treated as decision support. Their outputs require validation because historical schedules may contain bias, hidden constraints, poor estimates, or practices that are unsuitable for the current project. Sensitive commercial, personnel, or client data also requires appropriate security and privacy controls.

Good governance requires version control, approval authority, change logs, data-date discipline, and clear ownership of assumptions. A schedule should be explainable to the project team, sponsor, auditor, and—where relevant—contractual counterparties. Automation should make assumptions more visible, not less.

12. Limitations and Professional Judgment

CPM assumes that durations and relationships are sufficiently defined for deterministic calculation. It does not inherently model productivity loss, learning curves, weather, probabilistic failures, multiple resource skills, fatigue, political decisions, or quality rework. It also tends to represent logic as binary: an activity is either a predecessor or it is not. Real work may have partial releases, information quality thresholds, and iterative feedback.

The remedy is not to abandon networks but to supplement them. Use explicit assumptions, risk registers, probabilistic analysis where appropriate, resource-constrained scheduling, scenario analysis, and frequent reforecasting. A network is a model; models are useful precisely because their assumptions can be examined.

For resource loading and leveling, the central principle is sequential: validate scope and activity definition, validate logic, validate durations and calendars, calculate CPM, review criticality and constraints, then load resources and analyse capacity. If leveling changes relationships or dates, rerun the network and reassess the critical path. Resource decisions and schedule logic are interdependent, but they should not be confused.

Chapter Summary

Schedule and network planning provide the foundation for reliable resource analysis. A schedule transforms scope into time-phased activities connected by logical relationships. Activity-on-node networks support four dependency types: finish-to-start, start-to-start, finish-to-finish, and start-to-finish. Leads and lags express overlap or waiting but should be used transparently and sparingly.

CPM uses forward and backward passes to calculate earliest and latest dates. Total float measures allowable movement without delaying the relevant completion date, while free float measures movement without delaying an immediate successor. The critical path is conditional on the network, durations, calendars, constraints, and completion milestone. Near-critical paths, convergence points, resource limits, and uncertainty require active management.

Calendars affect dates and float; milestones provide control points; constraints represent external conditions; and baselines provide an approved reference for variance analysis. Before resource loading, planners should check logical completeness, relationship quality, duration realism, calendar consistency, constraints, criticality, progress integrity, and traceability. Software and AI can assist analysis, but professional judgment, governance, and transparent assumptions remain essential.

Key Terms

Activity: A defined piece of project work with measurable duration.

Activity-on-node: A network notation in which activities are represented by nodes and relationships by arrows.

Baseline: An approved schedule reference used for performance comparison.

Calendar: A definition of working and non-working time.

Critical path: The path governing the calculated completion date under stated assumptions.

Dependency: A logical relationship between activities.

Earliest finish: The earliest time an activity can finish according to network logic.

Earliest start: The earliest time an activity can begin according to network logic.

Float: Schedule flexibility available under specified conditions.

Free float: Delay possible without delaying an immediate successor.

Lag: Waiting or separation inserted into a logical relationship.

Lead: Overlap that permits a successor to proceed before a predecessor is fully complete.

Milestone: A zero-duration control or achievement point.

Network: Connected representation of activities and their logical relationships.

Predecessor: An activity that logically comes before another activity.

Resource leveling: Adjusting activity timing to resolve resource over-allocation.

Successor: An activity that logically follows another activity.

Total float: Delay possible without delaying the relevant project completion date.

Work breakdown structure: Hierarchical decomposition of project scope and deliverables.

Review Questions

5 Short Answer Questions

1. What is the difference between a WBS and a project schedule?
2. Define finish-to-start and finish-to-finish relationships.
3. What is the purpose of a calendar in network scheduling?
4. Distinguish total float from free float.
5. Why should a baseline not be changed merely to remove unfavorable variance?

5 Long Answer Questions

1. Explain how a precedence network is developed from scope and activity information.
2. Describe the forward-pass and backward-pass procedures in CPM, including the assumptions behind the standard equations.
3. Discuss the advantages and risks of using leads and lags in a project schedule.
4. Explain how milestones, constraints, and baselines support project governance.
5. Evaluate the limitations of deterministic CPM for complex, uncertain, and resource-constrained projects.

3 Analytical / Application Questions

1. Construct a precedence network for a project containing design, procurement, site preparation, installation, testing, and handover. Identify the relationships, likely milestones, and at least two schedule-quality checks.
2. Using the worked example in this chapter, assess how a two-day delay to activity C affects the critical path and project completion. State any assumptions clearly.
3. A project has a technically feasible network but one specialist is required simultaneously by two critical activities. Analyse the likely effects of resource leveling on dates, float, and criticality, and recommend a governance process for approving the revised schedule.

Multiple Choice Questions with Answers

Question 1

Which relationship means that a successor cannot start until its predecessor has finished?

A. Finish-to-start

B. Start-to-start

C. Finish-to-finish

D. Start-to-finish

Correct Answer: A. Finish-to-start

Question 2

In the worked CPM example, what is the calculated project duration?

A. 15 working days

B. 19 working days

C. 21 working days

D. 24 working days

Correct Answer: C. 21 working days

Question 3

Which activity is on the critical path in the worked example?

A. B Architecture

B. C Prototype

C. F Training

D. None of these

Correct Answer: B. C Prototype

Question 4

For a simple FS network, which expression calculates total float?

A. ES − LS

B. LS − ES

C. EF − LF

D. ES + duration

Correct Answer: B. LS − ES

Question 5

What does a positive lag generally represent?

A. An imposed reduction in duration

B. A waiting or separation period

C. An automatic resource increase

D. A deleted predecessor

Correct Answer: B. A waiting or separation period

Question 6

Which statement best describes a schedule baseline?

A. A continuously overwritten forecast

B. An approved reference for performance comparison

C. A list containing only milestones

D. A resource histogram without dates

Correct Answer: B. An approved reference for performance comparison

Question 7

What is the main danger of excessive hard constraints?

A. They always increase total float

B. They can mask the underlying network logic

C. They eliminate the need for calendars

D. They convert all activities into milestones

Correct Answer: B. They can mask the underlying network logic

Question 8

In the worked example, why does activity D start at day 9 rather than day 7?

A. Its duration is six days

B. Activity F is a predecessor

C. Activity C finishes later than activity B

D. The project calendar has a holiday

Correct Answer: C. Activity C finishes later than activity B

Question 9

Which check is most directly concerned with activities lacking predecessors or successors?

A. Logical completeness

B. Cost variance

C. Procurement compliance

D. Resource utilization

Correct Answer: A. Logical completeness

Question 10

Why should resource loading generally follow initial CPM and schedule-quality review?

A. Resources determine all logical relationships automatically

B. An invalid network can produce misleading resource requirements

C. CPM can be performed only after leveling

D. Baselines cannot contain resources

Correct Answer: B. An invalid network can produce misleading resource requirements