9. Scheduling and Operational Control

Introduction and context

Scheduling converts an operations plan into a time-based sequence of activities. It determines what will be done, when it will be done, where resources will be used, and in what order work will proceed. Operational control then compares actual performance with the schedule and initiates corrective action when deviations occur. Together, scheduling and control connect strategic capacity decisions with daily execution.

In logistics and operations, schedules influence customer service, inventory, labour utilisation, transport reliability, equipment usage and operating cost. A manufacturing schedule determines the release and completion of jobs. A warehouse schedule allocates receiving, picking, packing and dispatch activities. A transport schedule assigns vehicles, drivers, routes and delivery windows. In services, scheduling determines staff availability, appointment capacity and customer waiting time.

Scheduling is not simply the preparation of a timetable. It is a decision problem involving competing priorities, uncertain demand, limited capacity, set-up times, material availability, labour regulations and changing customer requirements. A theoretically efficient schedule may fail if it ignores congestion, interruptions, maintenance, safety or human factors. Effective operational control therefore requires both analytical methods and managerial judgement.

Learning objectives

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

1. Explain the role of scheduling and operational control in different operating environments.

2. Apply sequencing rules and interpret Gantt charts.

3. Distinguish infinite-capacity from finite-capacity scheduling.

4. Analyse line-balancing and workforce-scheduling decisions.

5. Construct and interpret CPM and PERT project schedules.

6. Select operational performance measures and control responses.

7. Evaluate digital scheduling tools, implementation risks and managerial trade-offs.

The scheduling problem: time, capacity and priority

A schedule allocates scarce resources to activities over time. Its basic elements are jobs or tasks, resources, processing times, release dates, due dates, precedence relationships and priorities. A job may represent a production order, a warehouse wave, a delivery trip, a customer appointment or a project activity.

Several objectives may conflict. A firm may seek to minimise completion time, average flow time, lateness, work-in-process inventory or overtime while maximising due-date adherence, utilisation and customer responsiveness. Consequently, there is no universally best scheduling rule. The appropriate method depends on the operating context and the relative cost of delay, idleness, overtime and inventory.

Important terminology includes:

• Processing time: Time required to perform an activity on a resource.
• Flow time: Elapsed time from a job’s release into the system until its completion.
• Due date: Required completion date or time.
• Lateness: Completion time minus due date; it may be negative or positive.
• Tardiness: The positive part of lateness, so early completion is not tardiness.
• Makespan: Time required to complete all jobs in a schedule.
• Utilisation: Proportion of available resource time that is productively used.
• Bottleneck: A resource or process stage whose limited capacity constrains system output.
• Slack: Time available before a job becomes late, often calculated as due date minus current time minus remaining processing time.
• Setup time: Time required to prepare a resource for a job, including cleaning, tooling or changeover.

A schedule should also be judged by feasibility. A feasible schedule respects resource capacity, precedence, material availability, working-time rules, maintenance windows and safety constraints. Optimisation without feasibility produces attractive but unusable plans.

Sequencing rules and dispatching logic

Sequencing determines the order in which waiting jobs are processed. Common rules include:

• First-come, first-served (FCFS): Jobs are handled in arrival order. It is simple and perceived as fair, but it may perform poorly when processing times differ substantially.
• Shortest processing time (SPT): The shortest job is processed first. SPT often reduces average flow time and work-in-process, although long jobs may wait excessively.
• Earliest due date (EDD): The job with the closest due date is selected first. It is useful when due-date performance is the principal objective.
• Critical ratio (CR): Critical ratio equals time remaining until due date divided by processing time remaining. A ratio below one indicates that the job is already at risk of delay. Jobs with the lowest ratios receive priority.
• Longest processing time (LPT): Longer jobs are scheduled first, often in parallel-machine or workload-balancing contexts.
• Slack per operation: Total slack is divided by the number of remaining operations. Lower values indicate greater urgency.

Sequencing rules may be static or dynamic. A static rule establishes an order once, whereas a dynamic dispatching rule is recalculated whenever conditions change. Warehouses and hospitals commonly require dynamic rules because arrivals, disruptions and priorities evolve during the operating day.

Rules also involve behavioural and ethical consequences. FCFS can support transparent service, while priority rules may disadvantage low-value or less influential customers. If a firm promises premium service, its priority policy should be explicit, consistently applied and compatible with contractual and regulatory obligations.

A simple sequencing illustration

Assume three jobs are available at time zero:

| Job | Processing time | Due date |

|---|---:|---:|

| A | 6 hours | 14 |

| B | 2 hours | 7 |

| C | 4 hours | 12 |

SPT produces B–C–A, with completion times 2, 6 and 12. EDD produces B–C–A because the due dates are ordered in the same way. If the due dates were A = 8, B = 20 and C = 12, EDD would produce A–C–B, while SPT would still produce B–C–A. The preferred sequence depends on whether the priority is average flow time or due-date performance.

No rule should be applied mechanically. A long job for a strategically important customer, a job requiring scarce material, or a job that releases work to a bottleneck may deserve priority even when a simple rule suggests otherwise.

Gantt charts and visual schedule control

A Gantt chart displays activities against time. Horizontal bars show when a job is planned to start and finish, while separate rows represent machines, work centres, vehicles, employees or project activities. Gantt charts are valuable because they make idle time, overlap, congestion and late work visible.

A basic Gantt chart is prepared by listing resources vertically, placing a time scale horizontally, assigning each job to the required resource and marking planned processing intervals. Actual progress can be shown with a different colour or line. The chart then becomes a control instrument rather than merely a planning document.

For example, if a packing station is scheduled to process Order X from 09:00 to 10:30 and Order Y from 10:30 to 12:00, a delay in X immediately threatens Y. The supervisor may add labour, move Y to another station, change the sequence or communicate a revised dispatch time. The chart supports rapid intervention, but it does not itself solve capacity or priority conflicts.

Gantt charts are particularly useful for repetitive production, warehouse waves, vehicle loading and maintenance planning. Their limitations become significant in complex projects with many precedence relationships, uncertain durations or frequent rescheduling. In such cases, network techniques and computer-based scheduling systems provide stronger analytical support.

Finite-capacity scheduling

Infinite-capacity scheduling assigns work without initially considering whether resources can actually perform it within the required period. It is useful for rough-cut planning and for identifying overall workload. However, it can release more work than a machine, dock, driver group or labour team can handle.

Finite-capacity scheduling explicitly respects available capacity. If a work centre has eight available hours and scheduled jobs require eleven hours, the excess work must be moved, subcontracted, assigned to another resource, performed as overtime or accepted as delay. Finite scheduling therefore exposes trade-offs rather than hiding them.

Capacity is not merely the number of machines. It includes skilled labour, dock doors, vehicle availability, storage locations, working hours, information systems and physical space. Effective capacity is lower than theoretical capacity because of breaks, maintenance, absenteeism, quality losses, changeovers and interruptions.

A practical finite-capacity process is:

1. Establish the planning horizon and time buckets.

2. Identify available resources and effective capacity in each bucket.

3. Load jobs according to material readiness, priorities and precedence.

4. Detect overloads and infeasible assignments.

5. Resolve conflicts through sequencing, reassignment, overtime, subcontracting, postponement or customer communication.

6. Freeze near-term commitments while retaining flexibility in later periods.

7. Monitor execution and replan when the operating environment changes.

The concept of a schedule freeze is important. Constantly changing near-term schedules creates instability, unnecessary setups and confusion. A frozen zone protects execution, while a flexible planning zone permits adjustment to new information.

Line balancing and flow-oriented scheduling

Line balancing assigns tasks among sequential workstations so that the required output is achieved with minimal idle time. It is relevant to assembly, sorting, order consolidation, cross-docking and other flow operations.

The cycle time required for a target output is:

`Cycle time = Available production time / Required output`

If 420 minutes are available in a shift and the required output is 70 units, cycle time is 6 minutes per unit. Each workstation must receive work whose total task time does not exceed six minutes. If total task time is 42 minutes, the theoretical minimum number of stations is 42/6 = 7, rounded up where necessary.

Balancing also requires precedence constraints. A task cannot be assigned before its predecessor has been completed. Common approaches include assigning the largest eligible task first, assigning the task with the greatest number of successors, or using optimisation models for larger problems.

Two useful measures are:

`Line efficiency = Total task time / (Number of stations × Cycle time) × 100`

`Balance delay = 100 − Line efficiency`

A high efficiency figure does not guarantee good performance. Excessively tight balancing may make the line vulnerable to small disruptions. Managers should consider ergonomic load, skill variation, quality checks, replenishment time and recovery capacity.

Workforce scheduling

Workforce scheduling assigns employees to shifts, locations and tasks while satisfying demand and employment constraints. Typical requirements include minimum staffing levels, skill qualifications, rest periods, maximum working hours, leave, fairness and labour agreements.

In a warehouse, the schedule may match labour to an hourly order profile. More pickers are required during the dispatch wave, while receiving labour may be concentrated earlier in the day. In a service centre, schedules must match agents to forecast contacts and required skills. A transport schedule must also consider driving-time regulations, breaks and fatigue risk.

Workforce scheduling can be approached manually for small operations, through spreadsheet models for moderate complexity, or through mixed-integer optimisation and specialised software for larger operations. The objective may minimise labour cost while maintaining service levels. Constraints should include not only contractual rules but also safe workload limits and reasonable predictability for employees.

A mathematically low-cost schedule can be operationally harmful if it produces unstable shifts, excessive split duties or chronic overtime. Good scheduling treats employee wellbeing, fairness and retention as operational variables rather than external concerns.

Project scheduling: CPM and PERT

Project scheduling is concerned with activities that occur once or a limited number of times and are linked by precedence relationships. Examples include warehouse construction, transport-network implementation, information-system deployment and facility relocation.

The Critical Path Method (CPM) uses normally known activity durations. A project network is constructed by identifying activities and dependencies. The forward pass determines earliest start and finish times; the backward pass determines latest start and finish times. The difference between latest and earliest timing is float. Activities with zero total float form the critical path, whose duration determines the minimum project completion time.

A critical path is not necessarily the most expensive path or the path containing the most activities. It is the path with no schedule flexibility under the current assumptions. As the project progresses, the critical path may change.

PERT is used when activity durations are uncertain. Three estimates are commonly obtained:

• optimistic time `a`;
• most likely time `m`;
• pessimistic time `b`.

Expected activity time is:

`te = (a + 4m + b) / 6`

The approximate activity variance is:

`Variance = [(b − a) / 6]²`

PERT helps managers assess completion risk, although its probability estimates depend on the quality of assumptions and the approximate treatment of dependencies. It should support judgement, not create false precision.

Project control includes milestone reviews, earned-value analysis, risk registers, change control and re-estimation of remaining work. A project that is on schedule in percentage terms may still be at risk if its remaining activities contain high uncertainty.

Operational control: from plan to response

Operational control is a closed-loop process. It begins with standards or targets, collects actual data, compares actual performance with the plan, diagnoses causes and applies corrective action. The loop is repeated because operations are dynamic.

A useful control sequence is:

1. Define the performance objective, such as on-time dispatch or vehicle utilisation.

2. Establish a measurable standard and reporting interval.

3. Capture reliable actual data.

4. Identify the size and significance of deviations.

5. Distinguish symptoms from root causes.

6. Select corrective or preventive action.

7. Confirm whether the action restored performance.

8. Update the schedule, standard or process when assumptions have changed.

Control charts can distinguish common-cause variation from special-cause variation. A recurring small deviation may indicate a flawed process standard, whereas a sudden major deviation may arise from equipment failure, a supplier problem or an unusual demand event. Treating every deviation as an individual worker failure encourages concealment and does not improve the system.

Key operational measures include schedule adherence, on-time-in-full delivery, throughput, cycle time, queue time, utilisation, changeover time, overtime, backlog, order accuracy, vehicle departure reliability and labour productivity. Measures should be balanced. Maximising utilisation at every resource can increase queues and reduce responsiveness, particularly when demand is variable.

The drum-buffer-rope perspective from the Theory of Constraints schedules the bottleneck as the drum, protects it with a buffer and controls release through the rope. This prevents excessive work-in-process from overwhelming the system. The central managerial insight is that improving a non-bottleneck does not necessarily improve total throughput.

Applied case: scheduling a regional fulfilment centre

Background

The following is an illustrative case. A regional fulfilment centre supplies retail outlets and e-commerce customers. Orders are picked in waves and must be packed before carrier collection at 18:00. The centre has two packing zones, a limited number of forklifts and a workforce whose availability varies by shift.

Situation and problem

Demand has increased, but late dispatches have also risen. Supervisors release all orders as soon as they arrive. The resulting workload is theoretically adequate for the day, yet the 16:00–18:00 period experiences severe congestion. Packers wait for replenishment, forklifts compete for narrow aisles and urgent orders interrupt planned work. Management initially proposes adding overtime to the final shift.

Analysis

A review of a two-week Gantt chart shows that the problem is not total daily capacity alone. It is the timing of release and the concentration of work near carrier cut-off. An infinite-capacity plan assumes that all orders can be processed in the desired windows. A finite-capacity analysis reveals that packing capacity between 16:00 and 18:00 is exceeded by 18 percent, while earlier periods contain idle time.

The centre classifies orders by carrier cut-off, service promise and handling requirements. It applies EDD within each service class, reserves capacity for genuinely urgent orders and uses a frozen release window for work already staged. A bottleneck analysis identifies packing replenishment, rather than picking, as the immediate constraint.

Approach and solution

The centre introduces four linked measures. First, waves are released earlier and more evenly, subject to inventory availability. Second, a finite-capacity schedule limits the number of pallets sent to packing during each hour. Third, one cross-trained employee is assigned to replenishment during the evening peak. Fourth, a visual control board displays planned versus actual wave completion, queue size and remaining time to carrier cut-off.

The centre does not eliminate all flexibility. A small protected capacity reserve is retained for late-arriving priority orders and recovery from disruptions. Supervisors review deviations every thirty minutes, but they do not repeatedly rewrite the entire shift schedule. This protects stability while allowing targeted intervention.

Outcome

The redesigned approach reduces peak congestion and makes late orders visible earlier. Overtime is used selectively rather than automatically. The most important improvement comes from aligning release timing with finite packing capacity, not from simply increasing labour at the end of the day.

Learning points

The case demonstrates that scheduling quality depends on the interaction of release decisions, bottlenecks, priorities and control intervals. It also illustrates why utilisation should not be interpreted in isolation. A resource can appear highly utilised while the overall system performs poorly because work is released too early or accumulated at a constrained stage.

Digital scheduling, analytics and emerging practice

Enterprise resource planning, warehouse-management, transport-management and workforce-management systems can integrate orders, inventory, capacity, routes and labour availability. Advanced planning and scheduling tools use constraint models, heuristics and optimisation algorithms to generate feasible schedules. Real-time dashboards compare planned and actual progress, while mobile scanning and telematics improve data accuracy.

Machine-learning models may forecast demand, estimate processing times, predict equipment failure or identify the probability of late delivery. However, predictive accuracy does not automatically produce a good schedule. Data may be incomplete, historical decisions may contain bias and an optimisation model may prioritise measurable cost over service, safety or fairness.

Human oversight remains essential. Managers should validate assumptions, establish approval rules for automated changes, preserve audit trails and define escalation procedures. A system that silently reassigns work, changes priorities or schedules excessive overtime creates governance risk. Cybersecurity, access control, data privacy and business continuity are also relevant where scheduling platforms are connected to operational equipment or employee information.

Emerging practice is moving toward event-driven scheduling: schedules are updated when a significant event occurs, such as a vehicle breakdown, order cancellation or inventory shortfall. Digital twins and simulation can test alternative policies before implementation. The danger is excessive rescheduling. Agility must be balanced with stability, because frequent changes can amplify disruption across suppliers, employees, carriers and customers.

Challenges, limitations and professional considerations

Scheduling models rely on assumptions. Processing times may be estimated incorrectly, demand may change, resources may be unavailable and precedence relationships may be misunderstood. A schedule can therefore be optimal for an inaccurate model. Managers should conduct sensitivity analysis and maintain contingency capacity for significant risks.

Common mistakes include:

• scheduling against theoretical rather than effective capacity;
• ignoring setup, travel, replenishment and handover time;
• releasing work before materials or information are available;
• optimising local utilisation while increasing system queues;
• treating due dates as priorities without considering customer impact;
• changing the schedule too frequently;
• measuring individual productivity without accounting for task difficulty;
• using automation without explainability or human escalation;
• concealing delays instead of reporting them early.

Professional control requires honest reporting of exceptions. Safety-critical work should never be expedited by bypassing inspections, rest requirements or competence standards. Scheduling decisions can also distribute undesirable shifts and workload unevenly. Transparent rules, documented exceptions and consultation with affected employees strengthen legitimacy and operational reliability.

Best-practice principles

Effective scheduling and control generally follow these principles:

1. Begin with the service promise and system constraint, not with a favourite sequencing rule.

2. Use finite capacity for execution decisions and rough-cut capacity for early planning.

3. Make priorities explicit and review them with relevant stakeholders.

4. Protect near-term stability while keeping later periods adjustable.

5. Integrate material, labour, equipment, space and transport constraints.

6. Measure both schedule adherence and the consequences of adherence, such as inventory or fatigue.

7. Use visual management for rapid local control and analytical systems for complex decisions.

8. Investigate root causes rather than blaming individuals for systemic variation.

9. Retain human judgement, auditability and ethical safeguards in automated scheduling.

10. Reassess the schedule when assumptions, bottlenecks or customer requirements change.

Chapter Summary

Scheduling allocates jobs, resources and priorities over time, while operational control compares planned and actual performance and coordinates corrective action. Sequencing rules such as FCFS, SPT, EDD and critical ratio support different objectives and must be selected contextually. Gantt charts provide practical visibility, whereas finite-capacity scheduling tests whether plans are executable. Line balancing aligns tasks with cycle time and precedence constraints; workforce scheduling combines demand, skills, labour rules and fairness. CPM identifies critical activities and float in projects, while PERT incorporates uncertain activity durations. Effective control requires reliable measures, root-cause analysis, disciplined rescheduling and balanced attention to throughput, service, cost, safety and employee wellbeing. Digital systems and AI can improve visibility and prediction, but they require sound data, governance and human oversight.

Key Terms

Scheduling: Allocation of activities and resources across time.

Sequencing: Determining the order in which jobs or tasks are processed.

Dispatching rule: A priority rule used to select the next job for processing.

Flow time: Time from a job’s release to its completion.

Lateness: Completion time minus the due date.

Tardiness: Positive lateness; delay beyond the due date.

Makespan: Completion time of the final job in a schedule.

Bottleneck: Capacity-constrained resource limiting system output.

Finite-capacity scheduling: Scheduling that explicitly respects available resource capacity.

Schedule freeze: A period during which near-term schedule changes are restricted.

Gantt chart: Time-based visual display of activities and resource assignments.

Cycle time: Available operating time divided by required output.

Line balancing: Assignment of tasks to workstations to meet output with minimal idle time.

Critical path: Longest precedence path determining minimum project duration.

Float: Time by which an activity may be delayed without delaying the project or successor constraint.

CPM: Network scheduling method using specified activity durations.

PERT: Probabilistic project-scheduling method using three time estimates.

Schedule adherence: Degree to which actual execution follows the planned schedule.

Operational control: Monitoring and corrective management of ongoing operations.

Drum-buffer-rope: Constraint-based method that schedules the bottleneck, protects it with a buffer and controls release.

Review Questions

Short Answer Questions

1. Distinguish lateness from tardiness.

2. What is the purpose of a Gantt chart?

3. Why can maximum utilisation be harmful to system performance?

4. Define finite-capacity scheduling.

5. What does float mean in a CPM network?

Long Answer Questions

1. Compare FCFS, SPT, EDD and critical-ratio sequencing rules, including the conditions under which each may be appropriate.

2. Explain how line balancing is performed and discuss the limitations of using line efficiency as the only performance measure.

3. Describe the stages of the operational-control loop and relate them to schedule deviations.

4. Explain CPM and PERT and compare their treatment of activity-time uncertainty.

5. Discuss how workforce scheduling can balance cost, service, compliance, fairness and employee wellbeing.

Analytical / Application Questions

1. Three jobs require 5, 3 and 7 hours and have due dates of 12, 9 and 15 hours respectively. Construct the SPT and EDD sequences and explain which objective each sequence supports.

2. A fulfilment centre has adequate daily capacity but experiences dispatch delays in the final two hours. Using finite-capacity scheduling and bottleneck logic, propose a diagnostic and corrective approach.

3. A project contains a chain of activities with uncertain durations and a promised completion date. Explain how PERT estimates, critical-path monitoring and contingency actions could be combined to manage completion risk.

22. Multiple Choice Questions with Answers

Question 1

Which sequencing rule generally minimises average flow time when all jobs are available and processing times are known?

A. First-come, first-served

B. Shortest processing time

C. Earliest due date

D. Longest processing time

Correct Answer: B. Shortest processing time

Question 2

A job is completed two days before its due date. What is its tardiness?

A. -2 days

B. 0 days

C. 2 days

D. It cannot be determined

Correct Answer: B. 0 days

Question 3

What distinguishes finite-capacity scheduling from infinite-capacity scheduling?

A. It ignores due dates

B. It schedules only projects

C. It explicitly respects available resource capacity

D. It eliminates the need for sequencing

Correct Answer: C. It explicitly respects available resource capacity

Question 4

A line has 420 available minutes and must produce 70 units. What is its cycle time?

A. 4 minutes

B. 6 minutes

C. 7 minutes

D. 70 minutes

Correct Answer: B. 6 minutes

Question 5

In CPM, activities on the critical path normally have:

A. The highest cost

B. The longest individual duration

C. Zero total float

D. The greatest number of workers

Correct Answer: C. Zero total float

Question 6

Which PERT estimate receives the greatest weighting in the expected-time formula?

A. Optimistic time

B. Most likely time

C. Pessimistic time

D. Actual completion time

Correct Answer: B. Most likely time

Question 7

A warehouse releases more work to packing than packing can process during the evening peak. Which response best reflects finite-capacity control?

A. Release all remaining work immediately

B. Ignore the overload because daily capacity is sufficient

C. Level releases, reassign capacity or revise priorities

D. Measure only picker utilisation

Correct Answer: C. Level releases, reassign capacity or revise priorities

Question 8

In the drum-buffer-rope approach, the ‘drum’ represents:

A. The customer due date

B. The bottleneck’s pace

C. The largest warehouse

D. The workforce schedule

Correct Answer: B. The bottleneck’s pace

Question 9

Why should automated scheduling systems retain human oversight?

A. Algorithms cannot display time periods

B. Scheduling decisions may involve bias, safety and judgement issues

C. Automation always reduces capacity

D. Human oversight removes the need for data

Correct Answer: B. Scheduling decisions may involve bias, safety and judgement issues

Question 10

A manager changes the near-term schedule repeatedly whenever a minor variation occurs. What is the most likely operational consequence?

A. Greater schedule stability

B. Lower setup and handover time

C. Instability and confusion during execution

D. Automatic elimination of bottlenecks

Correct Answer: C. Instability and confusion during execution