Detailed Scheduling and Sequencing

Introduction

Detailed scheduling converts a production plan into a time-phased, operational programme. It determines what should be processed, on which resource, in what order, and when work should start and finish. Whereas aggregate planning deals with product families and medium-term capacity, detailed scheduling operates at the level of individual jobs, orders, operations, machines, employees and project activities.

Scheduling decisions strongly influence customer delivery, machine utilisation, work-in-process (WIP), throughput, overtime, expediting and operating cost. A schedule is therefore not merely a timetable. It is a coordination mechanism linking customer requirements, material availability, labour, equipment, maintenance, quality controls and logistics.

Scheduling is especially difficult because production systems are constrained by limited capacity, variable processing times, breakdowns, absenteeism, rush orders, setup requirements and changing priorities. Good scheduling combines mathematical logic with managerial judgement and reliable shop-floor information.

Learning objectives

After studying this chapter, the learner should be able to:

1. Explain loading and distinguish finite from infinite capacity scheduling.
2. Apply forward and backward scheduling methods.
3. Construct and interpret Gantt charts.
4. Use common sequencing rules in job, flow and project environments.
5. Calculate flow time, lateness, tardiness, makespan, utilisation and WIP-related measures.
6. Compare scheduling approaches and identify their limitations.
7. Evaluate a practical schedule using operational and customer-service criteria.

1. The scheduling problem

A scheduling problem contains four basic elements:

• Jobs or activities: customer orders, batches, maintenance tasks or project activities.
• Resources: machines, work centres, tools, people, transport equipment or specialist teams.
• Processing requirements: the operations and time required by each job.
• Constraints: precedence, release dates, due dates, capacity, setup, material availability and labour rules.

A job may require several operations. For example, a fabricated component might require cutting, drilling, heat treatment, painting and inspection. The schedule must respect both the technological route and the availability of resources.

Important terminology includes:

• Release date: the earliest date or time at which a job is available for processing.
• Processing time: the time a job occupies a resource, excluding or including setup depending on the stated convention.
• Setup time: time required to prepare a machine or process for a job.
• Due date: the promised or required completion time.
• Completion time: the actual time at which processing is completed.
• Flow time: elapsed time from job release to completion.
• Lateness: signed deviation from the due date.
• Tardiness: positive lateness only; early completion has zero tardiness.
• Makespan: time from the beginning of the schedule to the completion of all jobs.
• Dispatching: selecting the next job to process when a resource becomes available.

For job j, let r_j be its release time, p_j its processing time, C_j its completion time and d_j its due date. The principal measures are:

where F_j is flow time. If all jobs are available at time zero, flow time equals completion time.

where L_j is lateness. A negative value means the job is early and a positive value means it is late.

where T_j is tardiness. These measures are applicable when completion and due dates use the same time units.

2. Loading: assigning work to capacity

Loading assigns work to machines, work centres, labour groups or project resources. It answers the question: “How much work is assigned to each resource, and during which period?” Loading may be performed before sequencing, although in practice the two activities are often interdependent.

2.1 Load calculations

For a resource, the gross required load over a planning period can be estimated as:

where L is required load, Q_i is the quantity of job or product i, t_i is processing time per unit, and s_i is setup time. All terms must use consistent time units.

Available productive capacity is often estimated as:

where H is scheduled hours, m is the number of equivalent resources, u is planned utilisation and e is expected efficiency. This is an approximation: utilisation and efficiency should not be multiplied blindly when they represent overlapping concepts.

A simple load ratio is:

A ratio above 1 indicates that the assigned load exceeds estimated available capacity. A ratio below 1 does not guarantee timely completion because priorities, interruptions and uneven operation routing may still create queues.

2.2 Loading approaches

Infinite loading assigns jobs without initially restricting the load to available capacity. It is useful for identifying the required workload, estimating overloads and revealing future bottlenecks. However, the resulting schedule may be impossible to execute.

Finite loading refuses to place work beyond the defined capacity of a resource or period. Excess work is moved to a later period, an alternative machine or an external supplier. Finite loading produces a feasible capacity-constrained schedule, but it may increase completion time or cause due-date violations.

Vertical loading gives priority to completing one order or job before loading the next. It is easy to understand but may leave some resources idle if the next job is not yet ready.

Horizontal loading distributes work across several orders or jobs, often by due date or priority. It can improve overall responsiveness but may increase setup frequency and WIP.

2.3 Capacity decisions

When overload is identified, managers may:

• reschedule work to another period;
• authorise overtime or additional shifts;
• use substitute equipment or qualified labour;
• subcontract selected operations;
• reduce setup time through preparation or SMED techniques;
• split a lot into smaller transfer batches;
• change priorities or negotiate due dates;
• redesign the process or remove non-value-adding work.

The decision should consider cost, quality, skill availability, safety, contractual obligations and the risk of creating a downstream bottleneck.

3. Forward and backward scheduling

3.1 Forward scheduling

Forward scheduling begins at the release date or the earliest feasible start. Each operation is placed as soon as the required resource and predecessor work are available. It answers: “If we start now, when can we finish?”

For a single uninterrupted operation with start time S_j and processing time p_j:

For a route with operations 1 to k, the start of operation i is constrained by both the completion of operation i − 1 and resource availability. Forward scheduling is useful for make-to-stock production, early completion, utilisation and identifying the earliest possible delivery date.

Its disadvantages are that it may create early finished goods, increase WIP and consume capacity before high-priority work becomes available.

3.2 Backward scheduling

Backward scheduling begins with the due date and works backwards. It answers: “When must we start to finish by the required date?” For a single operation:

where C_j is the required completion time, often the due date. For multiple operations, the latest finish of one operation becomes the latest start of its successor.

Backward scheduling reduces premature production and is valuable in just-in-time and make-to-order systems. It is vulnerable to disruption: if a machine fails or material arrives late, the schedule may immediately become infeasible. Practical systems therefore include protective time, capacity checks and controlled release rules.

3.3 Example: forward and backward scheduling

An order requires three sequential operations: cutting (2 hours), machining (4 hours) and inspection (1 hour). It is released at 08:00 on Monday and is due at 16:00 on Monday. Ignoring breaks and setup:

• Forward schedule: cutting 08:00–10:00, machining 10:00–14:00, inspection 14:00–15:00. The order finishes at 15:00.
• Backward schedule: inspection 15:00–16:00, machining 11:00–15:00, cutting 09:00–11:00. The planned start is 09:00.

The total processing time is 7 hours, leaving one hour of available time before the due date. The forward schedule finishes early, whereas the backward schedule protects against unnecessary early completion. Neither method alone accounts for competing jobs unless resource capacity and priorities are incorporated.

4. Gantt charts

A Gantt chart is a time-scaled chart showing jobs or activities as horizontal bars. The horizontal axis represents time, while the vertical axis lists machines, work centres, employees or activities. A bar’s position indicates start and finish; its length represents duration.

Gantt charts support:

• communication of the operating plan;
• identification of idle periods and overloads;
• monitoring of actual progress against plan;
• coordination of jobs across linked resources;
• rapid rescheduling after disruption.

A basic chart should distinguish planned and actual work, show maintenance and non-working periods, and identify critical milestones. It should be updated from reliable transactions rather than informal estimates.

Figure: A Gantt chart showing the timing of jobs on two machines.

The chart shows that Job A requires successive processing on Machine 1 and Machine 2. A schedule is feasible only if a machine is assigned to one operation at a time and technological precedence is respected.

5. Sequencing rules

Sequencing determines the order in which waiting jobs are processed. A rule should be selected according to the operating objective rather than applied automatically.

The critical ratio is commonly calculated as:

where d_j is the due date, t is the current time, and p_j^r is remaining processing time. A ratio below 1 indicates that the job is already behind its nominal plan if no allowance exists. Because the ratio changes as time passes, it is a dynamic dispatching measure.

5.1 Worked sequencing example

Three jobs are available at time zero on one machine:

Under SPT, the sequence is B–C–A. Completion times are 2, 6 and 12. The average flow time is:

Lateness values are B: 2 − 8 = −6; C: 6 − 10 = −4; A: 12 − 14 = −2. All jobs are early.

Under EDD, the sequence is B–C–A because due dates happen to have the same order. In other data, SPT and EDD can produce different results. SPT is often attractive when reducing average waiting is the main objective; EDD is more appropriate when protecting due-date performance is dominant.

The result depends on assumptions: all jobs are available, the machine processes one job at a time, processing times are known, pre-emption is not used, and there are no sequence-dependent setups. If these assumptions fail, the rule remains a heuristic rather than an optimisation guarantee.

6. Scheduling environments

6.1 Job-shop scheduling

In a job shop, different jobs follow different routes through shared resources. The environment offers high variety and relatively low volume. Scheduling is difficult because alternative routes, setups, rework and urgent orders generate complex interactions.

Typical approaches include dispatching rules, finite-capacity scheduling, constraint programming, mixed-integer optimisation and simulation. A practical job-shop system normally combines a stable master schedule with real-time dispatching.

6.2 Flow-shop scheduling

In a flow shop, jobs generally pass through machines in the same sequence. This makes the system more structured than a job shop. The main decisions concern job order, batch size, setup reduction and balancing.

For two machines, Johnson’s rule provides an optimal sequence for minimising makespan under restrictive assumptions: all jobs are available at time zero, each job visits Machine 1 then Machine 2, processing times are known and fixed, machines process one job at a time, no pre-emption occurs, and there are no significant sequence-dependent setups. The rule places the job with the smallest processing time among all remaining operations at either end: if it is on Machine 1, place the job as early as possible; if it is on Machine 2, place it as late as possible. For more machines, extensions require additional conditions or alternative methods.

6.3 Project scheduling

Projects consist of activities linked by precedence relationships and usually have a defined completion objective. Unlike repetitive production, a project may produce a unique output and involve changing resources.

The critical path method (CPM) uses known or relatively stable activity times. The critical path is the longest-duration path through the precedence network and determines the minimum project duration. Activities on the critical path have zero total float under the stated schedule.

For activity i, total float can be expressed as:

where ES_i and EF_i are earliest start and finish, while LS_i and LF_i are latest start and finish without delaying the project completion date.

When activity times are uncertain, programme evaluation and review technique (PERT) may be used. With optimistic time a, most likely time m and pessimistic time b, the conventional expected activity time is:

This is an estimate based on a beta-like weighting assumption, not a guarantee of actual duration. PERT variance is commonly estimated as:

The estimates should be treated cautiously when risks are correlated or when activity distributions are strongly non-symmetric.

7. Evaluating schedules

A schedule should be evaluated against the organisation’s objectives. Common measures include:

Makespan

C_max is the completion time of the last job. It is useful for measuring total schedule duration and is central in project and flow-shop scheduling.

Average completion and flow time

where n is the number of jobs. Lower average flow time generally indicates less waiting and WIP, assuming comparable release conditions.

Maximum lateness and total tardiness

Maximum lateness identifies the worst signed due-date deviation, while total tardiness focuses only on late work.

Utilisation

For a resource operating for T available hours with B busy hours:

High utilisation is not always desirable. In a variable system, operating at near 100% utilisation leaves little protection against variability and can cause queues to increase sharply.

Work-in-process and Little’s Law

For a stable system over a sufficiently long period:

where WIP is average work-in-process, TH is average throughput rate and CT is average cycle time. Units must be compatible; for example, units per day multiplied by days produces units. Little’s Law describes a relationship, not a scheduling rule. It assumes a stable system in which average inflow and outflow are approximately equal.

8. Integrated illustrative case: Riverside Components

Background

Riverside Components is an illustrative medium-sized manufacturer producing customised pump brackets. Orders pass through cutting, machining and inspection. The machining centre is the bottleneck, and customers have complained about late deliveries despite high machine utilisation.

Problem

At the start of a shift, four jobs are waiting at machining. All are available and the machine can process only one job at a time.

The supervisor has traditionally used FCFS: R1–R2–R3–R4.

Analysis

Under FCFS, completion times are 5, 7, 10 and 14. Lateness is −7, 1, 1 and 6 hours respectively. Two jobs are late, and the final job finishes at hour 14.

An EDD sequence is R2–R4–R3–R1, producing completion times 2, 6, 9 and 14. Lateness is −4, −2, 0 and 2. The maximum lateness falls from 6 to 2 hours, and the contract-critical job is completed first.

The SPT sequence is R2–R3–R4–R1, with completion times 2, 5, 9 and 14. Lateness is −4, −4, 1 and 2. This produces a lower average completion time than FCFS, but R4 is completed later than under EDD.

Approach and solution

Riverside selects EDD as the primary rule because customer due-date reliability is the immediate objective. It adds a controlled priority override for contract-critical work, verifies material availability before release, and uses a finite-capacity schedule for machining. A small time allowance is reserved for tool changes and unexpected minor stoppages rather than loading the machine to its theoretical maximum.

The supervisor reviews the schedule at fixed intervals instead of changing priorities after every telephone call. Emergency overrides require a recorded reason and an assessment of their effect on other due dates.

Outcome

The revised sequence reduces worst-case lateness and improves visibility of the contract-critical order. It does not reduce makespan in this example because the total processing time is unchanged. Management therefore avoids claiming that a sequencing rule alone increases capacity. Subsequent improvement work focuses on reducing setup time and improving upstream release discipline.

Learning points

1. The best rule depends on the performance objective.
2. A schedule can improve due-date performance without reducing makespan.
3. Finite capacity prevents infeasible promises.
4. Priority overrides require governance to prevent arbitrary expediting.
5. Bottleneck improvement may be more valuable than changing sequence rules repeatedly.

9. Technology, implementation and governance

Modern manufacturing execution systems, enterprise resource planning systems and advanced planning and scheduling tools can calculate loads, simulate alternatives, issue dispatch lists and update schedules from shop-floor data. Artificial intelligence and machine learning may assist with demand prediction, estimated processing times, anomaly detection and dispatching recommendations.

However, technology does not eliminate the need for sound process design. Poor master data, inaccurate routings, unrealistic processing times and missing downtime records will produce apparently sophisticated but unreliable schedules. AI recommendations should be explainable enough for planners to understand the trade-offs, and human approval remains important where safety, contractual commitments or employment rules are involved.

Good governance includes:

• clear ownership of schedule changes;
• version control and an audit trail;
• approved priority categories;
• protection of confidential customer and employee information;
• compliance with working-time, safety and quality regulations;
• separation of planned, released and completed status;
• periodic review of forecast errors and schedule adherence.

A practical implementation sequence is to establish accurate routings and calendars, identify bottlenecks, select meaningful performance measures, pilot a finite-capacity schedule, train supervisors, and then integrate real-time feedback. Schedule adherence should be distinguished from schedule quality: a poor schedule can be followed perfectly, while a good schedule may require legitimate revision after a breakdown.

10. Challenges and limitations

Scheduling models commonly assume known processing times, stable resources and accurate due dates. Real operations violate these assumptions. Important sources of uncertainty include:

• machine failure and maintenance;
• variable quality and rework;
• material shortages;
• absenteeism and skill constraints;
• sequence-dependent setup times;
• rush orders and engineering changes;
• transport and batch-transfer delays;
• inaccurate inventory or job-status data.

No single sequencing rule optimises every objective. Minimising average flow time can disadvantage long jobs; maximising utilisation can increase WIP; strict EDD can create excessive setups; and frequent rescheduling can confuse operators and destabilise material planning. Managers should therefore use a hierarchy of objectives, such as safety and regulatory compliance first, then contractual due dates, quality, flow and cost.

Best-practice checklist

1. Define the objective before selecting a rule.
2. Confirm release dates, routings, processing times and resource calendars.
3. Identify bottlenecks and protect them from avoidable starvation.
4. Use finite capacity where customer promises depend on feasibility.
5. Separate planned, released and completed work.
6. Include setup, maintenance, breaks and realistic allowances.
7. Use Gantt charts for communication, not as a substitute for analysis.
8. Review schedule performance using flow time, tardiness, makespan, WIP and utilisation together.
9. Establish controlled procedures for priority overrides.
10. Replan when conditions change, but avoid unnecessary nervousness in the schedule.

Chapter Summary

Detailed scheduling translates production requirements into an executable sequence of jobs and activities. Loading assigns work to resources, while finite loading explicitly respects capacity and infinite loading reveals workload requirements without initially enforcing capacity limits. Forward scheduling starts at the earliest feasible time; backward scheduling starts from the due date and works towards the required release time.

Gantt charts communicate timing, resource occupancy and progress. Sequencing rules such as FCFS, SPT, EDD, critical ratio and slack time serve different objectives and must be applied with awareness of their assumptions. Job shops require flexible dispatching and capacity coordination; flow shops benefit from structured sequencing; and project environments use precedence networks, CPM and, where times are uncertain, PERT estimates.

Makespan, flow time, lateness, tardiness, utilisation and WIP provide complementary measures of schedule performance. Little’s Law links WIP, throughput and cycle time in stable systems. Effective scheduling combines quantitative methods, accurate data, finite-capacity thinking, controlled managerial judgement and continuous improvement.

Key Terms

Loading: Assigning work requirements to resources and time periods.

Finite-capacity scheduling: Scheduling that does not assign work beyond available resource capacity.

Infinite-capacity scheduling: Scheduling that initially assigns work without enforcing capacity limits.

Forward scheduling: Planning from the earliest feasible start towards completion.

Backward scheduling: Planning backwards from a required completion or due date.

Gantt chart: A time-scaled visual display of jobs or activities.

Sequencing: Determining the order in which waiting jobs are processed.

Dispatching: Selecting the next job for a resource.

Processing time: Time required to perform an operation.

Setup time: Time required to prepare a resource for an operation.

Flow time: Elapsed time from job release to completion.

Lateness: Signed difference between completion time and due date.

Tardiness: Positive lateness; early completion is counted as zero.

Makespan: Completion time of the final job or activity.

Work-in-process: Jobs started but not yet completed.

Critical path: Longest precedence path determining minimum project duration.

Total float: Time an activity can be delayed without delaying project completion.

Critical ratio: Remaining time until due date divided by remaining processing time.

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

Review Questions

Short Answer Questions

1. What is the difference between finite and infinite loading?
2. Define flow time, lateness and tardiness.
3. State one advantage and one limitation of backward scheduling.
4. What information is normally represented in a Gantt chart?
5. Why can high machine utilisation increase WIP?

Long Answer Questions

1. Explain the role of loading in detailed production scheduling and discuss methods for responding to an overloaded work centre.
2. Compare forward and backward scheduling, including their assumptions, benefits and operational risks.
3. Explain FCFS, SPT, EDD and critical-ratio sequencing rules and identify an appropriate objective for each.
4. Discuss how scheduling differs among job-shop, flow-shop and project environments.
5. Explain how makespan, flow time, tardiness, utilisation and WIP should be used together to evaluate a schedule.

Analytical / Application Questions

1. Four jobs have processing times of 3, 7, 2 and 5 hours and due dates of 9, 12, 6 and 15 hours respectively. Construct SPT and EDD sequences, calculate each job’s completion time and tardiness, and recommend a sequence for a due-date-focused operation.
2. A work centre has 40 scheduled hours, planned utilisation of 0.85 and expected efficiency of 0.90. Its assigned work consists of 30 units requiring 0.8 hours each plus 6 hours of setup. Calculate the estimated available capacity and load ratio, then interpret the result.
3. A project contains activities A (3 days), B (4 days after A), C (2 days after A), and D (3 days after both B and C). Draw the precedence logic, identify the critical path and calculate the minimum project duration.

Multiple Choice Questions with Answers

Question 1

Which scheduling approach explicitly prevents work from being assigned beyond available resource capacity?

A. Infinite-capacity scheduling

B. Finite-capacity scheduling

C. Forward loading only

D. Random dispatching

Correct Answer: B. Finite-capacity scheduling

Question 2

A job is released at time 2 and completed at time 11. What is its flow time?

A. 2 time units

B. 9 time units

C. 11 time units

D. 13 time units

Correct Answer: B. 9 time units

Question 3

Which sequencing rule generally minimises average flow time on a single machine when all jobs are available and assumptions are satisfied?

A. Shortest processing time

B. Earliest due date

C. First come, first served

D. Longest processing time

Correct Answer: A. Shortest processing time

Question 4

A job is completed at time 18 and has a due date of 15. What is its lateness?

A. −3 time units

B. 0 time units

C. 3 time units

D. 33 time units

Correct Answer: C. 3 time units

Question 5

What is the principal purpose of backward scheduling?

A. To start every job immediately

B. To determine the latest feasible start time for meeting a due date

C. To maximise machine idle time

D. To ignore precedence relationships

Correct Answer: B. To determine the latest feasible start time for meeting a due date

Question 6

A resource is busy for 32 hours during 40 available hours. What is its utilisation?

A. 0.80 or 80%

B. 1.25 or 125%

C. 8%

D. 72%

Correct Answer: A. 0.80 or 80%

Question 7

Which measure represents the completion time of the final job in a schedule?

A. Average flow time

B. Tardiness

C. Makespan

D. Slack time

Correct Answer: C. Makespan

Question 8

In a project network, the critical path is the path that has which characteristic?

A. The fewest activities

B. The longest total duration

C. The greatest positive float

D. The lowest resource requirement

Correct Answer: B. The longest total duration

Question 9

Using Little’s Law, if average throughput is 10 units per day and average cycle time is 3 days, what is average WIP?

A. 3.33 units

B. 7 units

C. 13 units

D. 30 units

Correct Answer: D. 30 units

Question 10

For a two-machine flow shop under its restrictive assumptions, which rule is designed to minimise makespan?

A. Johnson’s rule

B. FCFS only

C. Random selection

D. Critical-ratio rule

Correct Answer: A. Johnson’s rule