SYE6047 Quality Improvement: Systems and Methodologies
The Complete Exam Bible · Semester A 2026/27SYE6047 Overview
- City University of Hong Kong
- Semester A 2026/27
- Postgraduate course
- Engineering
- Two hour sample exam
What this course asks you to connect
SYE6047 joins management systems, improvement methods and statistical evidence. It begins by treating quality as fitness for a stated or implied need, then separates the dimensions along which a product or service can succeed or fail.
- Assessment split The group presentation carries 20%, the group report 30% and the final examination 50%.
- Exam pattern The sample paper has a multiple-choice section and a long-question section with all questions required.
- Core method Translate customer language into a CTQ, build a baseline, validate causes, prove a remedy and specify control.
- Hardest distinction Statistical control describes process behaviour; capability compares that behaviour with specifications.
How SYE6047 is assessed
| Component | Weight | Format |
|---|---|---|
| Group Project Presentation | 20% | Presentation of the selected quality improvement case |
| Group Project Report | 30% | Complete group report submitted as a PDF on Canvas |
| Final Examination | 50% | Sample paper: two hours, closed book, 70% multiple choice and 30% long questions |
The published course outline lists these three components and the numeric weights total 100%. The sample examination gives the examination format used in this guide. Confirm the live examination and submission arrangements on Canvas.
Assessment structure
Segment widths reproduce the published weights and total 100%.
Current SYE6047 dates
| Date | Item | Control |
|---|---|---|
| 30 September 2026 | Group formation | Group formation deadline in the project instruction. |
| 22 December 2026 | Group Project Report | Report deadline at 11:59 p.m. in the project instruction. |
Dates are as published in the group project instruction for Semester A, 2026/27. Confirm exact deadlines and submission settings in the live LMS.
What SYE6047 covers
The chapters follow the published teaching sequence from quality definitions and improvement systems through Six Sigma, DMAIC and quality tools to process control, measurement systems and statistical decisions.
Meaning of Quality and Eight Dimensions
customer requirements, three quality lenses, Garvin dimensions and evidence based trade-offsQuality begins with use
Quality is not a decorative synonym for excellence. It is the degree to which a product or service satisfies stated and implied needs.
- Quality as satisfaction of stated and implied needs
- Product, manufacturing and user lenses
- Performance and features
- Reliability and conformance
- Durability and serviceability
- Aesthetics and perceived quality
Quality Paradigms and Improvement Thinkers
paradigm evolution, Deming, Juran, Crosby, Feigenbaum and the cost of preventionResponsibility moves upstream
The history of quality can be read as an expansion of responsibility. Inspection separates unacceptable output after production. Statistical quality control detects and reduces variation during production.
- Inspection and detection
- Statistical quality control
- Quality assurance
- Total quality management
- Deming's system view
- Juran's quality trilogy
Six Sigma Variation and Defect Metrics
variation, DPU, DPMO, specification width and potential process capabilityVariation is the operating problem
Six Sigma links customer requirements to the distribution of process output. The mean describes location and the standard deviation describes spread, but neither alone establishes performance.
- Location and spread
- Common and special variation
- Units, defects and defectives
- Opportunities per unit
- DPU calculation
- DPMO calculation
Define and Measure in DMAIC
project charters, VOC to CTQ translation, SIPOC, operational definitions and sampling plansDefine the gap without solving it
A DMAIC project begins with an observable gap, not a favourite cause or solution. A useful problem statement identifies the symptom, magnitude, location and period.
- Problem statement
- Business case and goal statement
- Project scope
- Voice of customer
- Critical-to-quality translation
- SIPOC boundaries
Analyze, Improve and Control in DMAIC
cause validation, root cause analysis, FMEA, action planning and control plansAnalysis converts possibilities into evidence
The Analyze phase asks which inputs or process conditions actually explain the verified performance gap.
- Candidate and validated causes
- Root cause analysis
- Cause evidence
- Solution generation
- Pilot design
- Failure mode analysis
Basic Quality Improvement Tools
check sheets, histograms, Pareto, fishbone, scatter, control charts and line graphsTools answer different questions
The seven basic quality tools are useful because each gives structure to a particular evidence problem. A check sheet creates consistent observations. A histogram displays the distribution of a quantitative variable.
- Check sheets and operational categories
- Histograms and distribution shape
- Pareto bars and cumulative share
- Cause-and-effect diagrams
- Scatter diagrams
- Control charts
New Quality Tools and Project Scheduling
affinity, relations, tree, matrix, arrow, PDPC and matrix data analysisStructure verbal complexity
The new seven quality tools are most useful when the evidence is qualitative, relationships are tangled or implementation must be planned.
- Affinity grouping
- Relations and influence
- Tree decomposition
- Matrix relationships
- Arrow networks
- Critical path
Variable Control Charts and Capability
x-bar and R limits, signal interpretation, baseline revision and CpTwo charts answer two linked questions
When rational subgroups contain measurements, the x-bar chart monitors subgroup average and the R chart monitors within-subgroup range.
- Rational subgrouping
- Subgroup means and ranges
- Grand mean and average range
- R-chart limits
- X-bar chart limits
- Signal investigation
Attribute Control Charts and Revision
p-chart limits, assignable causes, revised baselines and process interpretationMonitor a fraction in time order
A p chart follows the fraction of inspected units that are nonconforming in each sample. It is appropriate when each unit can be classified into two states under a consistent definition and sample sizes are known.
- Binary unit classification
- Sample fraction nonconforming
- Pooled centre line
- Equal-size p-chart limits
- Unequal-size limits
- Signal investigation
Measurement Systems and Hypothesis Testing
bias, stability, linearity, Gage R and R, test tails and evidence based decisionsObserved variation has two sources
An observed measurement combines variation in the item or process with variation introduced by the measurement system.
- Measurement system components
- Bias
- Stability
- Linearity
- Repeatability
- Reproducibility
That vocabulary matters because an improvement action must target the dimension the customer actually values rather than a convenient proxy.
The course then follows the expansion of quality responsibility from inspection through statistical control and assurance to organisation-wide improvement.
From Six Sigma language to a DMAIC project
Six Sigma supplies measures of variation, defects and opportunities as well as a project infrastructure.
The guide keeps the calculations tied to operational definitions so a lower DPMO cannot be manufactured by inflating opportunity counts.
DMAIC provides the main project route: define one observable gap, translate voice of customer into a measurable critical-to-quality characteristic, plan a trustworthy baseline, test competing causes, prove a remedy and build a control plan that survives handover.
Tool choice before calculation
The basic and new quality tools answer different questions.
Check sheets structure collection, histograms show distributions, Pareto diagrams rank categories, fishbones organise possible causes, scatter diagrams show association and control charts monitor behaviour in time. Affinity, relations, tree, matrix, arrow, process decision and matrix data tools organise verbal evidence and implementation.
The guide always names what a visual can support and what it cannot prove.
Where the quantitative work concentrates
The later chapters build x-bar and R charts, p charts, process capability ratios and measurement-system diagnostics. Worked examples show every substitution and finish with an interpretation.
Measurement system analysis separates bias, stability, linearity, repeatability and reproducibility before hypothesis testing is used to judge evidence of a difference. The practice chapter uses original numbers while following the question types and marking moves visible in the published sample examination.
How to revise efficiently
First learn the boundaries between closely related terms.
Next recompute each worked example without looking at the answer. Finally add the decision sentence: whether a process is stable, whether capability has been assessed, whether a cause is validated and what evidence comes next. This order prepares both the recognition demanded by multiple-choice questions and the structured calculations demanded by long questions.
Worked example · free
Build and interpret an attribute control chart
- 4Calculate the average fraction and three-sigma limits from the original samples.
- 3Identify the signal and require an assignable-cause investigation before exclusion.
- 5Recompute transparently after a verified removed cause and interpret the revised process.
Key terms
- Critical To Quality
- A measurable product or process requirement translated from a customer need and used to focus improvement work.
- Common Cause
- Variation generated by the current process system and expected while that system remains unchanged.
- Special Cause
- A detectable source of variation linked to a particular event or process change that warrants investigation.
- DPU
- Defects per unit, calculated as the observed number of defects divided by the number of units processed.
- DPMO
- Defects per million opportunities, a scaled rate that divides defects by units times defined opportunities per unit.
- Process Capability
- The ability of a stable process distribution to fit within customer or design specification limits.
- Control Limit
- A process-derived boundary showing the expected range of a plotted statistic under current common-cause variation.
- Specification Limit
- A boundary set by design or customer requirements rather than calculated from process behaviour.
- Repeatability
- Variation when one appraiser repeatedly measures the same characteristic with the same instrument and method.
- Reproducibility
- Variation between appraisers' average measurements of the same characteristic using the same instrument and method.
- Measurement Bias
- The difference between the average observed measurement and the accepted reference value.
- Linearity
- Change in measurement bias across the expected operating range of an instrument.
- Null Hypothesis
- The statistical claim tested against a stated alternative using a chosen significance level and decision rule.
SYE6047 FAQ
How is the course assessed?
The published outline assigns 20% to the Group Project Presentation, 30% to the Group Project Report and 50% to the Final Examination. These rows total 100%. The project instruction gives group and submission arrangements, while the sample examination provides the section structure used for revision in this guide.
What format does the sample examination use?
The sample paper allows two hours, is closed book and permits an HKEA-approved calculator. It contains 35 multiple-choice questions and two long questions, with every question required. Section A carries 70 marks and Section B 30 marks. Confirm the current examination notice before the actual sitting.
What is the difference between control and capability?
Statistical control asks whether process behaviour is consistent with common-cause variation over time. Capability asks whether a stable process distribution fits within specification limits. A process can be stable but incapable, or appear within specifications while unstable. Read the variation chart and mean chart before comparing the process with customer tolerances.
How should a DMAIC project begin?
Begin with one observable problem statement that contains a symptom, magnitude, location and period without inserting a cause or solution. Translate the customer issue into a measurable CTQ, set project scope with SIPOC and develop an operational definition before collecting a baseline. This prevents later analysis from answering a different question.
Which quality tool should I choose first?
Choose from the decision required. Use a check sheet to structure counts, a histogram for a distribution, Pareto for priority, fishbone for possible causes, scatter for association and a control chart for behaviour over time. Affinity diagrams organise unstructured verbal data before a relations diagram proposes causal links.
When can an out-of-control point be removed?
A point beyond a control limit triggers investigation. Exclude it from a revised baseline only when an assignable cause is verified, the affected data are identified consistently and the cause has been removed or separated from future operation. Preserve the original chart. Deleting a point merely because it is extreme biases the limits.
What does measurement system analysis add?
Measurement system analysis separates process variation from error introduced by the standard, workpiece, instrument, appraiser, procedure and environment. It distinguishes bias, stability, linearity, repeatability and reproducibility. Without that diagnosis, a team can adjust a stable process in response to gauge noise or misclassify products near specification limits.
Questions students ask in each SYE6047 chapter
- Can a product conform and still have poor quality?
- Yes. It can meet every drawing tolerance yet fail the user's purpose because the specification captured the wrong need. Conformance is essential evidence for a manufacturing claim, but it does not replace the user lens. Revisit the customer, intended use and critical requirement before celebrating consistent output against a target that may be irrelevant. Record the revised need as a measurable characteristic before changing the design target. Chapter 01 · Meaning of Quality and Eight Dimensions
- Did total quality management eliminate inspection?
- No. It changed the role of inspection by placing it inside a wider prevention and learning system. Detection remains useful where risk warrants it, but inspection alone cannot improve the process that creates defects. Its findings need feedback to design, suppliers and process owners so recurring failures become inputs to prevention. The test is whether detection changes the system that produced the result. Chapter 02 · Quality Paradigms and Improvement Thinkers
- Why are defects and defective units different?
- One invoice can contain several errors but is only one defective invoice. DPU counts all defects, whereas a fraction defective counts units with at least one defect. Mixing them changes both numerator meaning and statistical model. Write the unit and counting rule before calculating so repeated defects on one item cannot silently become several defective items. Preserve both raw counts when a problem reports them. Chapter 03 · Six Sigma Variation and Defect Metrics
- Why should a problem statement exclude causes?
- An untested cause narrows investigation and makes contrary evidence harder to see. Define the performance gap first; candidate causes belong in Analyze after the measurement system and baseline are credible. A neutral statement also lets the team compare explanations and remedies without treating disagreement as resistance to a solution already embedded in the charter. It keeps project approval separate from solution approval. Chapter 04 · Define and Measure in DMAIC
- Is a fishbone diagram proof of root cause?
- No. It organises possible causes and supports questioning. Validation requires evidence that differentiates the candidate from alternatives and connects it to the measured process outcome. Translate a branch into a prediction, compare it with rival explanations and retain it only when the observed pattern or a controlled change supports the proposed mechanism. Document rejected branches so they do not return during handover without new evidence. Chapter 05 · Analyze, Improve and Control in DMAIC
- When should I use a histogram instead of a bar chart?
- Use a histogram for ranges of one quantitative variable, where adjacent bins represent a scale. Use bars for distinct categories. Sorting histogram bins by height would destroy the distribution's meaning. Preserve bin boundaries and units so shape, spread and unusual gaps remain interpretable, while category bars may be ordered to support a separate priority question. State the observation period on either display. Chapter 06 · Basic Quality Improvement Tools
- How is an affinity diagram different from a fishbone diagram?
- Affinity grouping lets categories emerge from many statements. A fishbone uses a cause-oriented structure to organise possible explanations for one effect. The former is useful before the problem structure is clear. Moving too quickly to cause branches can discard observations that do not fit the team's first theory and make tentative influence look established. Chapter 07 · New Quality Tools and Project Scheduling
- Why inspect the R chart first?
- X-bar limits depend on an estimate of within-subgroup variation. If ranges are unstable, that estimate is not dependable, so interpretation of the average chart can be misleading. Investigate range signals and establish a defensible spread baseline before treating mean-chart limits as evidence of process location. Chapter 08 · Variable Control Charts and Capability
- Why do p-chart limits change with sample size?
- The standard error of a sample fraction is inversely related to the square root of sample size. Smaller samples vary more naturally, so their limits are wider when the pooled fraction is unchanged. Carry the inspected total beside every point and calculate its own boundary; an average denominator can hide a signal or manufacture one. Chapter 09 · Attribute Control Charts and Revision
- Can calibration at one reference prove linearity?
- No. It can estimate bias near that reference. Linearity requires reference values across the intended operating range and analysis of whether bias changes with magnitude. Stability also needs repeated checks over time, while reproducibility needs measurements across appraisers or other stated conditions. Select references that span the actual decision range, not merely values that are convenient to measure, and preserve the same documented method across the comparison. Chapter 10 · Measurement Systems and Hypothesis Testing
How to study for the exam
Build revision around contrasts and decision chains. First write one-line pairs: control limits versus specifications, repeatability versus reproducibility, common versus special cause, affinity versus relations, and stability versus capability. Next cover every worked result and reproduce the formula, substitution and final number.
Add a sentence that states what the result establishes and another that names the remaining assumption. For control-chart practice, follow the published sample's order: calculate the initial centre and limits, plot, identify signals, investigate assignable causes, revise only with evidence, and interpret the revised process. For multiple choice, eliminate options that place a tool in the wrong phase or reverse a direction.
For the group project, use the same route on the chosen case, but keep customer need, measurement plan, causal evidence, remedy proof and control ownership visible as separate decisions. Before the examination, practise without notes and use only the calculator functions you expect to rely on under the published closed-book conditions.
