How to perform in the new G2 SEC Mathematics examination at an advanced level includes knowing when a mathematically correct number is not yet a valid answer. Real problems contain constraints: capacity, budget, whole-number counts, minimum requirements, maximum limits, valid input ranges, geometry conditions and physical boundaries. The learner must solve the mathematics and then test whether the result is feasible.
This thirty-fifth Learner’s Guide focuses on constraints and feasibility. The central rule is: before accepting a result, ask what conditions the answer must satisfy. A decimal bus, negative length, probability above 1 or value outside a graph’s valid range may come from correct arithmetic applied without enough interpretation.
For 2027, SEAB lists G2 Mathematics as K210 under the Singapore–Cambridge Secondary Education Certificate. Use the SEAB 2027 G2 syllabus page for official year-specific details and the Complete Mathematics Index for canonical topic learning.
A Constraint Is Part of the Mathematics
Students sometimes treat conditions in the wording as background. They are often part of the mathematical model.
“At least 20”, “no more than $50”, “whole number of tickets”, “positive length”, “within the interval”, “must fit inside” and “cannot exceed capacity” all restrict the acceptable solution set.
The Constraint-First Read
Before calculating, mark words that limit the answer:
- at least;
- at most;
- more than;
- less than;
- minimum;
- maximum;
- whole number;
- positive;
- between;
- within;
- not exceed;
- must fit;
- available budget or time.
These words should influence the model from the start, not appear only during checking.
At Least
“At least” includes the boundary. If at least 30 students are needed, 30 is valid. In inequality language, the quantity is greater than or equal to the boundary.
In practical counting problems, a calculated value below 30 is not feasible even if the arithmetic itself is correct.
At Most
“At most” also includes the boundary. If a budget is at most $120, exactly $120 is acceptable.
The learner should distinguish this from “less than $120”, which excludes the boundary.
More Than and Less Than
These strict inequalities exclude the boundary. The language difference is small but mathematically important.
During checking, substitute the boundary verbally: does the wording allow exactly this value?
Minimum and Maximum
A minimum can refer to the smallest feasible count, length, time, cost or other quantity. A maximum refers to the largest.
The calculated mathematical threshold may need interpretation before becoming the final discrete answer.
Whole-Number Constraints
People, buses, boxes, tickets, chairs and many other real-world quantities must be whole numbers. A calculation producing 4.2 buses does not mean 4.2 buses are possible.
The correct practical answer may require rounding up or down depending on the condition. The direction of rounding must be justified by the constraint, not by ordinary rounding rules alone.
The Ceiling Decision
If 4.2 buses are required to carry everyone and partial buses are impossible, five buses are needed. The reason is capacity, not because 4.2 rounds to 4 under standard rounding.
Always explain practical rounding with the real constraint.
The Floor Decision
If a problem asks how many complete packages can be made from a fixed supply, a decimal result may need to be rounded down because incomplete packages do not count.
Again, the direction comes from the situation.
Positive Quantities
Lengths, masses, times and many counts cannot be negative in ordinary contexts. An algebraic solution may produce a negative root that is mathematically valid for the equation but invalid for the physical quantity.
Do not reject a negative solution automatically. Reject it because the context or domain makes it impossible.
Probability Boundaries
Probabilities lie between 0 and 1 inclusive when expressed in decimal form. A result outside this range is immediate evidence of a setup or calculation problem.
Use the boundary as a built-in feasibility check.
Angle Boundaries
Angles in geometric contexts must satisfy the properties of the figure. A triangle angle cannot be negative, and the set of angles must fit the relevant angle-sum relationships.
The diagram and geometry provide constraints beyond the equation alone.
Graph Domains
A graph or function may be meaningful only over a certain input range. A mathematical extension beyond the domain may not answer the real problem.
If a graph represents time during a two-hour experiment, interpreting it at five hours is unjustified unless the question explicitly asks for extrapolation and the model supports it.
Data-Range Constraints
Statistical and graphical conclusions should remain inside the observed range unless a justified prediction is requested.
The fact that a line continues visually does not guarantee that the real-world relationship continues indefinitely.
Capacity Constraints
Capacity problems require the learner to compare demand with a maximum available amount: seats, storage, volume, weight limit, bandwidth or another bounded resource.
A solution is feasible only if the constraint is satisfied after all units and practical rounding are handled.
Budget Constraints
Budget questions are not complete when the cost is calculated. Compare the cost with the allowed budget and state whether the plan is feasible.
If multiple options exist, the cheapest option may still be infeasible if it exceeds another condition.
Time Constraints
Schedules can impose start times, deadlines, duration limits or available intervals. A travel calculation may be correct but useless if arrival occurs after the required time.
Translate clock times carefully and return the duration result to the schedule context.
Geometry Fit Constraints
Objects must fit inside shapes or spaces. A diameter, diagonal, height or width may create a limiting dimension even when area or volume appears sufficient.
The learner should identify which geometric dimension actually controls feasibility.
Multiple Constraints
Advanced real-world problems may have several constraints at once: budget, time, capacity and whole-number counts.
A solution must satisfy all of them. Passing one constraint does not compensate for failing another.
The Constraint Table
For complex problems, create a short table:
- constraint;
- mathematical condition;
- calculated result;
- pass/fail.
This is particularly useful when comparing several plans or options.
Feasibility Is a Final Mathematical Operation
After solving, add one final operation: interpret. The learner should be able to state whether the numerical solution actually works in the situation.
This final operation is often short, but it distinguishes a calculation from a complete model.
The Feasibility Sentence
Practise writing one sentence after selected real-world questions: “This is feasible because…”, “This is not feasible because…”, or “The minimum whole number required is…”
The sentence forces the learner to connect the result to the condition.
Inequalities as Constraint Language
Inequalities are compact representations of verbal constraints. Translate both directions.
- x ≥ 12 → x is at least 12;
- x ≤ 50 → x is at most 50;
- x > 7 → x is more than 7;
- x < 3 → x is less than 3.
The learner should be able to test whether the boundary is included by reading the original words.
Compound Constraints
Some quantities must satisfy two boundaries simultaneously: for example, 10 ≤ x ≤ 25.
In context, this means the value must be at least 10 and at most 25. A result outside either boundary fails.
Integer Domains
If x represents a number of people, tickets or objects, the valid domain may be integers even if the algebra allows all real numbers.
State or remember the domain before solving. This prevents accepting a decimal solution that the context does not allow.
Non-Negative Domains
Some quantities can be zero but not negative: distance travelled, elapsed time, number of items or mass in ordinary contexts.
A mathematical root outside the domain can be rejected after solving, provided the reason is stated or understood.
Restricted Input Domains
Functions involving denominators, square roots or contextual limits can have restricted inputs. The learner should know where the expression is defined and where the context is meaningful.
Do not substitute values outside the valid domain simply because the calculator returns something.
Constraint Reading in Word Problems
Long word problems often contain the decisive constraint in a small phrase near the end: “each bus holds at most…”, “the total must not exceed…”, “only complete sets can be sold”.
Read the final line twice before accepting the answer.
Constraint Reading in Tables
A table may include maximum capacities, available quantities or thresholds. Treat those values as conditions, not just data.
When comparing options, check every required column rather than choosing based on one attractive figure.
Constraint Reading in Graphs
Graphs can impose visual bounds: allowed interval, intersection threshold, maximum height, minimum value or a region where a condition is satisfied.
Shade or mark the valid region during practice when this helps make the constraint visible.
Constraint Reading in Geometry
Geometry constraints can include positive lengths, angle limits, triangle conditions, parallelism, congruence, similarity and shape fit.
An algebraic solution that violates the geometry is not acceptable.
The Triangle Feasibility Check
Three positive lengths do not automatically form a triangle. The learner should know the relevant side relationship when the syllabus context requires it.
Feasibility can therefore depend on geometric properties beyond positivity.
The Area-versus-Dimension Trap
A shape can have enough area but still fail to fit because one dimension is too long. Real-world fit problems require dimensional checking, not only total area comparison.
Ask which dimension is the actual bottleneck.
Capacity and Rounding
Capacity questions often require rounding away from ordinary nearest-value conventions. If every person needs a seat, 4.1 buses means five buses. If only complete packages count, 4.9 packages may mean four complete packages.
The context determines the direction.
Budget and Rounding
Money may be rounded to the relevant currency precision, but feasibility should use the correct unrounded or appropriately rounded comparison depending on the task.
Do not round a cost downward in a way that makes an unaffordable option appear affordable.
Time and Schedule Feasibility
A calculated duration must be added to the correct start time and compared with the deadline or opening hours.
Crossing an hour, noon or midnight can create interpretation errors even when the duration calculation is correct.
Rate Constraints
Rates may have maximum or minimum limits: speed limits, production capacity or resource-use rates. A calculated rate should be compared with the allowable range.
A model may also assume a constant rate. If the context makes that assumption unreasonable, note the limitation if the question asks for judgement.
Percentage Constraints
Percentages in many contexts lie within expected ranges, but not all percentages are automatically limited to 100%. Percentage increase, for example, can exceed 100%.
Do not apply a boundary rule mechanically. Use the meaning of the quantity.
Probability Is Different
Probability itself is bounded by 0 and 1. This is a mathematical definition, not a contextual convention.
A result outside the range is a strong check that the setup or arithmetic is wrong.
Negative Roots and Context
Quadratic equations may produce two roots. The context may allow both, one or neither.
Test each root against the variable meaning and all stated constraints. Do not reject a root merely because it is negative if the variable can legitimately be negative, and do not keep it if the quantity cannot.
Extraneous Solutions
Some algebraic manipulations can introduce solutions that do not satisfy the original equation or domain. Substitute important final solutions back into the original relationship.
Feasibility checking begins with mathematical validity before real-world validity.
The Two-Stage Validity Check
- Stage 1: Does the result satisfy the mathematics?
- Stage 2: Does the result satisfy the context?
An answer must pass both.
Feasibility in Optimisation
When comparing options, the best numerical value may still violate another condition. A lowest cost option may take too long; a fastest option may exceed budget.
Optimisation is usually “best subject to constraints”, not simply “smallest” or “largest”.
The Feasible-Set Idea
Think of all solutions satisfying the constraints as the feasible set. Only solutions inside this set are candidates for the final decision.
This idea helps when several inequalities or conditions act together.
Constraint Errors Are Often Reading Errors
A learner may know the Mathematics but lose marks because one condition was never translated. The repair is not more algebra; it is constraint marking and final interpretation.
Keep constraint errors separate in the error ledger.
The Constraint Error Ledger
- boundary inclusion misread;
- whole-number requirement missed;
- negative contextual value accepted;
- unit limit ignored;
- budget or time condition not checked;
- graph domain exceeded;
- geometry condition violated;
- wrong practical rounding direction;
- one of several constraints forgotten;
- mathematically valid result not translated into a feasible decision.
These categories are highly trainable.
The Constraint-Only Drill
Take ten real-world questions and do not solve them. Underline every constraint and translate it into mathematical language.
This isolates reading and modelling before calculation.
The Feasibility-Only Drill
Provide completed calculations, some feasible and some not. The learner decides whether each final answer is valid in context and explains why.
This trains interpretation without the distraction of solving.
The Rounding-Direction Drill
Use count, capacity and packaging problems where ordinary rounding would give the wrong practical answer. Require a sentence explaining the direction.
This breaks the habit of applying nearest-number rounding mechanically.
The Multiple-Constraint Drill
Give three options with cost, time and capacity. Ask which options satisfy all conditions before asking which is best.
The learner practises feasibility before optimisation.
The Domain Drill
Use equations with several mathematical solutions and give each variable a context. Decide which solutions belong to the valid domain.
This connects algebra to meaning.
The Graph-Region Drill
Use a graph with a threshold line or allowed interval. Identify the region satisfying the condition, then translate it into inequality language.
Move back and forth between visual and symbolic constraints.
The 20-Minute Constraint Session
- Five minutes: identify constraints in five problems.
- Ten minutes: solve two selected problems.
- Five minutes: perform mathematical-validity and contextual-feasibility checks.
This can be integrated into Paper 2 preparation.
A Four-Week Constraint Build
Week 1 — language
At least, at most, strict inequalities, whole-number conditions and domains.
Week 2 — real-world feasibility
Capacity, budget, time and practical rounding.
Week 3 — geometry and graphs
Valid regions, dimensions, angle and shape constraints.
Week 4 — mixed timed transfer
Use real-world Paper 2 questions with multiple constraints and require final feasibility statements.
Constraints and Representation Switching
Use Vol 0031. Verbal constraints can become inequalities, graph regions, table conditions or diagram labels.
Constraints and Checking
Use Vol 0019. Feasibility is one layer of reasonableness checking.
Constraints and Mathematical Communication
Use Vol 0023. State the practical conclusion, not only the calculated number.
Use Examination Craft
For timing, return decisions and checking, continue through the Examination Craft hub. Constraint reading belongs at the start; feasibility checking belongs at the end.
The PSLE Bridge
The PSLE habit Represent Before You Calculate remains important. At G2, representation should include not only relationships but limits and valid domains.
Final Rule
A numerical answer is not complete until it passes the conditions of the problem.
Read the constraints before solving. Preserve them in the model. Test mathematical validity. Test contextual feasibility. Round in the direction the real situation requires. The final answer must be possible, not merely calculable.
Scenario 1 — The Bus Capacity Problem
A school needs transport for 173 students. Each bus carries at most 40 students. Dividing gives 4.325 buses. Standard rounding to the nearest whole number would give four, but four buses carry only 160 students.
The feasibility condition is capacity. The answer must be five buses. The mathematical division finds the threshold; the constraint decides the practical integer.
Scenario 2 — Complete Packages
A supplier has enough material for 7.8 complete packages. Only complete packages can be sold. The answer is seven complete packages, not eight.
The same decimal structure as the bus problem produces the opposite rounding direction because the constraint changed.
Scenario 3 — Budget and Quality
Two plans fit a budget, but only one meets the minimum quality requirement. Choosing the cheaper plan without checking the second constraint is incomplete optimisation.
Feasibility must be tested across all stated conditions before ranking feasible options.
Scenario 4 — Deadline and Travel Time
A journey takes 1.75 hours and begins at 2:40 p.m. The learner must translate 0.75 hours into 45 minutes, calculate the arrival time and compare it with the deadline.
A correct duration that is never returned to clock time does not complete the problem.
Scenario 5 — Geometry Fit
A rectangular object has area smaller than the area of a storage space but one side is longer than the available width. The object does not fit in the stated orientation.
Area is not the controlling constraint. Dimension is.
Scenario 6 — Negative Root
An equation for a length produces roots 5 and -3. Both may satisfy the algebraic equation, but the negative value is not feasible as a length in the stated context.
The learner should reject -3 because of the domain, not because negative answers are always wrong.
Scenario 7 — Probability Above One
A probability calculation produces 1.2. No contextual interpretation can rescue it. The value violates a defining mathematical boundary.
Return to event counting, denominator choice or arithmetic before continuing.
Scenario 8 — Graph Extrapolation
A graph shows a relationship up to x = 20. The question asks about x = 18: interpolation is inside the observed region. A later question asks about x = 50: the learner must recognise that this is extrapolation and that the model may be less reliable.
The graph domain is part of the judgement.
Scenario 9 — Minimum Integer Solution
An inequality yields x > 6.4 and x represents the number of boxes. The smallest valid integer is 7.
The learner must combine inequality meaning with integer domain.
Scenario 10 — Maximum Integer Solution
A budget inequality yields x ≤ 12.8 and x is a whole-number quantity. The maximum feasible integer is 12, assuming no other condition changes the interpretation.
The algebra gives a bound; the domain converts the bound into a practical answer.
Scenario 11 — Percentage More Than 100%
A quantity grows from 40 to 100. The percentage increase is 150%. This is valid because percentage increase can exceed 100%.
Do not apply the probability boundary to percentages simply because both use percent notation.
Scenario 12 — Time Cannot Be Negative but Coordinates Can
A negative time duration is usually infeasible in ordinary travel problems, while a negative coordinate or temperature may be perfectly meaningful.
Feasibility depends on the variable’s meaning, not the sign alone.
Scenario 13 — Measurement Precision
A calculation gives 12.34678 cm, but the input measurements and question require a stated degree of accuracy. The final answer should use the required precision without pretending the measurement is more exact than the data supports.
Accuracy conventions are part of valid communication.
Scenario 14 — Resource Allocation
A project requires at least 25 units of resource A and at most 40 units of resource B. A proposed plan satisfies the first condition but uses 42 units of B.
The plan is infeasible because every constraint must hold simultaneously.
Scenario 15 — Ratio Constraint
A mixture must keep a ratio within a stated range. The total quantity alone is not enough. The learner must test the composition as well as the amount.
This is another example of multiple constraints acting on one solution.
The Constraint Stack
For a complex problem, list constraints from most structural to most practical:
- mathematical domain;
- equation or inequality conditions;
- units and dimensions;
- whole-number or discreteness condition;
- real-world capacity, time or budget;
- final rounding and communication.
A solution should pass each layer before being accepted.
The Boundary Test
When an inequality or threshold appears, test the boundary explicitly.
- Is the boundary included?
- What happens just below it?
- What happens just above it?
- Does the context require an integer nearest to the boundary?
This prevents confusion between strict and non-strict conditions.
The Domain Test
Before solving, state what values the variable is allowed to take: real numbers, positive values, non-negative values, integers or a contextual interval.
After solving, remove solutions outside the domain.
The Feasibility Test
After solving, ask whether the result can exist physically or operationally. Can a person count be fractional? Can the object fit? Can the event occur within the time? Can the plan stay within budget?
The test converts abstract result into practical answer.
The Optimisation Test
If choosing the best option, first filter out infeasible options. Then compare the objective among the remaining options.
This avoids selecting a numerically attractive option that violates a hidden requirement.
The Constraint-to-Inequality Drill
Give ten verbal conditions and translate them into inequalities without solving any full problem.
Then reverse the process: read the inequality and state the condition in words.
The Feasibility Audit Drill
Provide five completed solutions. Some are mathematically correct but contextually invalid. The learner identifies the exact failed constraint.
This separates calculation skill from interpretation skill.
The Domain-and-Root Drill
Use equations with multiple roots and assign different variable meanings. Decide which roots are valid in each context.
The same algebra can produce different accepted solutions when the domain changes.
The Multiple-Constraint Option Drill
Present three options with cost, time, capacity and quality data. Require a pass/fail table before selecting the best option.
This trains feasibility filtering before optimisation.
The Graph-Boundary Drill
Use graphs with thresholds and valid regions. Ask the learner to shade or identify where a condition is satisfied, then express the region as an inequality.
This builds visual-symbolic constraint transfer.
The Rounding-Reason Drill
For every whole-number result, the learner must explain why the answer rounds up, rounds down or remains exact.
“Because normal rounding says so” is not accepted unless ordinary rounding is actually the relevant rule.
The Constraint Check Under Time
Late in the paper, use a compact final check:
- domain;
- boundary;
- unit;
- whole-number condition;
- real-world feasibility.
This is often enough to catch a mathematically neat but invalid answer.
Constraint Errors in Paper 1
Short questions may hide one key word such as “maximum”, “minimum” or “integer”. The learner should mark these words during the first read.
Because Paper 1 moves quickly, constraint reading must become automatic.
Constraint Errors in Paper 2
Longer Paper 2 problems may contain several interacting conditions. Use a quick list or table before modelling.
The extra structure is justified because one forgotten condition can invalidate a long calculation chain.
Constraint Errors and Error Containment
Use Vol 0027. If a constraint was omitted, locate the earliest point where the model stopped respecting it and repair from there.
Constraint Errors and Mark Security
Use Vol 0025. Constraint words often protect accessible marks. Missing “at least” or “whole number” can turn a correct calculation into a lost mark.
Constraint Errors and Representation
Use Vol 0031. Constraints can be represented as inequalities, graph regions, table limits or diagram boundaries.
Advanced Standard: Feasible Mathematics
The advanced learner does not stop at “I got a number”. They know what the number represents, what values were allowed, which boundaries matter and whether the result can exist in the real situation.
Feasibility is the point where mathematical technique becomes mathematical judgement.
Final Perspective
Constraints are not obstacles added after the Mathematics. They define the problem.
Read them early, represent them explicitly, preserve them through the calculation and test them again at the end. A valid answer must satisfy both the mathematics and the world the mathematics is modelling.
The Final Feasibility Checklist
- Does the solution satisfy the original equation or relationship?
- Is it inside the allowed domain?
- Are boundary conditions interpreted correctly?
- Does the unit match the quantity?
- Does the context require a whole number?
- If rounding is needed, does the direction follow the practical constraint?
- Does the result satisfy every budget, time, capacity or geometry condition?
- If several options remain, have infeasible options been removed before comparison?
This checklist is most useful on long Paper 2 problems and real-world applications. The learner does not need to write every item; the habit should become a rapid mental scan.
The Constraint-Conflict Scenario
Some problems contain competing constraints. The cheapest plan may be too slow. The fastest route may exceed budget. The largest capacity may not fit the available space. In these cases, there may be no solution satisfying every condition, or the best solution may be a compromise inside the feasible set.
The learner should never assume that one option must work simply because the question presents several choices. Test feasibility first.
The Constraint-Communication Rule
When the final answer depends on a practical constraint, state the reason. “Five buses are required because four buses carry only 160 students” is stronger than a bare “5”. “Twelve complete packages can be made” communicates the discrete limit better than writing only the calculator output.
Mathematical communication makes the feasibility decision visible and checkable.
Final Standard
The advanced G2 learner treats conditions as part of the model from the first read. They translate wording into boundaries, preserve those boundaries through working and reject answers that fail either mathematical validity or contextual feasibility.
The final number is only a candidate. The problem decides whether the candidate is allowed.
The Final Constraint Calibration
Constraint errors should be reviewed separately from calculation errors. A learner who repeatedly solves correctly but ignores whole-number conditions does not need more arithmetic practice. They need a stronger constraint-first read and a feasibility sentence at the end.
Likewise, a learner who misreads “at least” and “more than” needs boundary-language practice, while a learner who accepts negative physical quantities needs domain interpretation. The repair should match the failed constraint.
The Mixed-Constraint Finish
Advanced Paper 2 problems may combine discrete counts, inequality boundaries, budget, time and geometry. The learner should list the conditions before optimising. A mathematically attractive option that fails one condition is removed from consideration before the remaining feasible options are ranked.
This “filter, then optimise” habit prevents long calculations from being built around an impossible candidate.
The Final Feasibility Standard
A complete G2 Mathematics answer passes three gates: the calculation is valid, the value belongs to the allowed mathematical domain and the result satisfies the real-world conditions of the question. Only then is the number ready to become the final answer.
When one gate fails, repair that layer rather than assuming the entire method is wrong. Constraint control is the discipline that keeps correct Mathematics connected to a possible world.
One Last Feasibility Rule
When several constraints appear together, do not optimise before filtering. First remove every option that violates a domain, boundary, whole-number, budget, time, capacity or geometry condition. Only then compare the remaining feasible options.
This order protects the learner from spending time proving that an impossible option is numerically attractive. The best answer must first be allowed to exist.
That is the final discipline of constrained Mathematics: solve, filter, interpret, then decide.
Constraint control should also survive unfamiliar wording. Whether the condition appears as an inequality, a sentence, a graph boundary, a capacity table or a geometric limit, the learner should recognise the same mathematical job: define what values are allowed, solve inside that space and reject every candidate that falls outside it.
In the final check, the learner should be able to name the constraint that makes the answer valid. If no such condition can be stated, return to the wording, domain or practical model before accepting the result.
Feasibility is complete only when the final value survives every stated condition.
