Constraint-based scheduling explained for audit planners
Constraint-based scheduling in plain terms: hard and soft rules, how an engine searches for the best compliant audit plan, and what planners still decide.
Key takeaways
- Constraint-based scheduling separates rules that can never break from goals that can be traded off.
- Hard constraints in audit planning come from standards, accreditation and scheme rules; soft goals come from your business.
- A deterministic engine gives the same answer for the same data, which makes every allocation explainable.
- Planners still set the priorities, decide exceptions and approve the plan.
Constraint-based scheduling is a method that builds a plan by first ruling out every option that breaks a hard rule, such as an unqualified auditor or a date outside the audit window, and then choosing among the remaining options the one that best meets soft goals, such as less travel or more internal work. Planners review and approve the result.
What is constraint-based scheduling?
Constraint-based scheduling is how most serious scheduling engines work, from airline crew rostering to hospital shifts. You describe the rules a valid plan must obey and the goals a good plan should achieve. The engine then searches the possible plans for one that obeys every rule and scores best on the goals.
For an audit planner, this is the same reasoning you already do in your head: first, who is allowed to do this audit; then, of those, who is the best choice. The difference is scale. An engine can weigh every auditor against every audit in the programme at once, which a person with a spreadsheet cannot.
A condition that a schedule must satisfy (a hard constraint) or should satisfy as far as possible (a soft constraint). In audit planning, competence for the scope is hard; keeping travel low is soft.
Which audit planning rules are hard and which are soft?
Getting this split right is most of the work. If a real rule is treated as soft, the engine may break it when the trade-off looks attractive. If a preference is treated as hard, the engine may report that no plan exists.
| Constraint | Type | Where it comes from |
|---|---|---|
| Auditor competent for standard and technical area on the audit date | Hard | ISO/IEC 17021-1 and your competence criteria |
| Auditor approved under the accreditation body for that standard | Hard | Accreditation scope |
| No conflict of interest with the client | Hard | Impartiality requirements |
| Audit inside its window or before its due date | Hard | Certification cycle and scheme rules |
| Auditor available and not double booked | Hard | Calendars and holidays |
| Rotation of auditors who have visited the client before | Hard or soft, depending on your policy | Scheme rules and your procedures |
| Minimise travel distance and overnight stays | Soft | Cost and emissions goals |
| Prefer internal auditors over subcontractors | Soft | Margin and utilisation goals |
| Client's preferred dates and language | Soft, or hard where required | Client requests and audit needs |
Rotation is a good example of a policy choice. Some schemes set explicit limits; elsewhere it is your procedure. Our guide to auditor rotation rules covers the detail.
How does a constraint-based scheduling engine find a plan?
The engine does not guess. It works through a defined sequence, and each step can be inspected afterwards.
- 1Load the dataAudits, auditors, competence, calendars, locations
- 2Filter by hard rulesRemove every invalid auditor and date
- 3Score the goalsWeigh travel, internal share, preferences
- 4Search combinationsCompares whole plans at once
- 5Explain the resultReasons and rejections for each allocation
The fourth step is where software earns its keep. Choosing the nearest qualified auditor for each audit in turn, which is what people do, often leaves later audits with no valid option. Searching across the whole programme avoids that trap. Our article on how an allocation engine decides goes further.
A worked example: one week, four auditors
Imagine a two-day ISO 14001 surveillance audit for a client in the North West, due this week. Four auditors exist in the pool. The scene below shows what the engine sees once hard constraints are applied.
- MonAuditor BHolidayUnavailable
- TueAuditor AConflictConsulted for client
- TueAuditor CBest optionQualified, 40 km
- WedAuditor AConflictConsulted for client
- WedAuditor CBest optionQualified, nearby audit Thu
- WedAuditor DNot qualifiedCode in training
- ThuAuditor BAvailableQualified, 280 km
- FriAuditor BAvailableQualified, 280 km
✓Result: Auditor C, Tue to Wed, with a nearby audit clustered on Thursday
Auditor A is ruled out by a conflict of interest, and Auditor D is still in training for the code, so neither is considered however convenient. Between B and C, soft goals decide: C is closer and can continue to another audit nearby. If C later falls ill, the engine re-runs and proposes B, showing the extra travel so the planner can decide.
Why does deterministic constraint-based scheduling matter for accreditation?
A deterministic engine gives the same answer every time it receives the same data. That sounds minor, but it is what makes a plan auditable. You can reproduce any allocation, see which rules were checked, and show an assessor why a particular auditor was chosen.
Engines that produce different results on each run, or that cannot say why they chose someone, are hard to defend in an accreditation assessment. Our article on an audit trail for scheduling decisions explains what assessors typically look for.
What do planners still decide?
Constraint-based scheduling moves the mechanical checking into software. The judgement stays with people, and a good engine makes that judgement easier to exercise.
- Which rules are hard and which are soft, and how soft goals are weighted
- How to handle a case where no valid plan exists, such as a shortage of qualified auditors
- Whether to accept a reschedule that adds cost to protect a client relationship
- Approval of every allocation before it reaches auditors and clients
See the planner's role with AI scheduling for how the job changes in practice.
What goes wrong with constraint-based scheduling?
Most failures come from data and set-up, rarely from the engine itself.
MythThe engine will work out our rules.
RealityIt applies the rules you give it. Unwritten rules, such as a client who only accepts one auditor, must be captured as data or they will be ignored.
MythIf no plan is found, the engine is broken.
RealityUsually a hard constraint is too strict or a data gap makes an auditor look unqualified. A good engine tells you which rule blocked each audit.
MythOptimisation means the cheapest plan.
RealityIt means the best plan by the goals you set. Weight internal utilisation or client preference higher and the plan changes accordingly.
How do you prepare data for constraint-based scheduling?
An engine is only as good as the competence and calendar data behind it. Before a trial, check these basics.
- ✓Competence recorded by standard, technical area and role, with expiry dates
- ✓Accreditation body per standard and per auditor where relevant
- ✓Previous auditors per client for at least one certification cycle
- ✓Known conflicts of interest recorded against auditor and client
- ✓Audit windows or due dates on every audit
- ✓Auditor home locations and client site addresses
Our competence matrix guide and the audit scheduling glossary help standardise the terms before you start.
ScheduleAI is the audit scheduling software certification bodies use to plan ISO programmes from stage 1 to recertification.
ScheduleAI uses a deterministic optimisation engine that applies 35+ scheduling parameters as hard and soft constraints, so the same data always gives the same, explainable plan, and planners approve every change.
Book a demo Estimate your savingsQuestions
What is constraint-based scheduling in simple terms?
Describing the rules a plan must obey and the goals it should meet, then letting software search for the plan that obeys all the rules and scores best on the goals.
What is the difference between a hard and a soft constraint?
A hard constraint can never be broken, such as auditor competence for the scope. A soft constraint is a goal traded against others, such as lower travel.
Is constraint-based scheduling the same as AI?
It is a branch of optimisation often grouped under AI. Unlike a language model, a deterministic constraint engine gives the same answer for the same input and can explain each decision.
Why does an engine sometimes find no solution?
Because at least one audit has no option meeting every hard rule. The usual causes are a genuine shortage of qualified auditors or missing competence data.
Can planners override the engine?
Yes. Planners approve every allocation and can change any of them. The engine then shows the effect of the change on rules, travel and other audits.