Decision support systems and structured, semi-structured and unstructured decisions: HSC Enterprise Computing Intelligent Systems
“Investigate applications of decision support systems; describe categories of decision-making within decision support systems, including an infinite set of instructions where judgement may be required (unstructured), use of a specified set of finite instructions (semi-structured) and automated decisions (structured)”
A decision support system combines data, models and an interactive interface to help people decide, in fields from healthcare to agriculture. Structured decisions follow clear rules and can be automated; semi-structured decisions follow a finite set of instructions but need judgement; unstructured decisions have no set procedure and rely on judgement supported by the DSS.
Jump to a section
What this dot point is asking
You need to investigate where decision support systems (DSSs) are used and describe the three categories of decision-making the syllabus names. Note NESA's wording: unstructured decisions involve an effectively unlimited set of possibilities where judgement may be required; semi-structured decisions use a specified, finite set of instructions; structured decisions are automated.
The answer
What a DSS is
A decision support system is an interactive information system that helps people make decisions by combining:
- Data management: internal data (sales, stock) and external data (weather, market prices).
- A model base: statistical models, forecasts, optimisation and what-if analysis.
- A user interface: dashboards, charts and scenario controls.
- Users: the decision makers, who interpret results and decide.
A DSS supports rather than replaces human judgement.
Applications of DSSs
- Healthcare: staffing forecasts, clinical decision support that flags drug interactions.
- Agriculture: when to plant, irrigate or harvest using weather, soil moisture and price data.
- Finance: loan assessment, investment portfolio scenarios.
- Logistics: route planning and fleet scheduling.
- Retail: pricing, promotions and stock ordering.
- Government and emergency services: bushfire spread modelling, resource allocation.
Categories of decision-making
- Structured (automated decisions): routine, repetitive, clear rules and complete information. The system can decide automatically (reorder stock at the reorder point, calculate pay, approve a small refund).
- Semi-structured (a specified set of finite instructions): some steps follow set rules, but judgement is needed to complete the decision (a loan that meets most criteria, choosing between suppliers, scheduling staff).
- Unstructured (judgement required): new, complex or one-off decisions with many possibilities and no set procedure (launching a new product, responding to a crisis). A DSS provides information and scenarios, but people decide.
As decisions move from structured to unstructured, human judgement matters more and automation is less appropriate. DSSs add the most value for semi-structured and unstructured decisions, where people need analysis but also judgement.
A farm management DSS supports three decisions.
- Structured: an irrigation controller turns on water when soil moisture falls below 30%. Automated.
- Semi-structured: choosing when to harvest. The DSS applies set rules on crop maturity and forecast rain, then the farmer weighs contractor availability and market prices.
- Unstructured: whether to convert part of the farm to a different crop in response to climate trends. The DSS models yields and prices under scenarios, but the farmer decides.
- Calling a DSS an expert system
- A DSS supports decisions with data and models; an expert system reasons with encoded expert knowledge (rules) to reach conclusions.
- Mixing up the categories
- Ask how much of the decision can follow fixed rules.
- Saying the DSS makes the decision
- For semi-structured and unstructured decisions, people decide.
Practice questions
Original practice questions graded from foundation to exam level, each with a full worked solution. Try them before revealing the solution.
foundation3 marksClassify each decision as structured, semi-structured or unstructured: (a) calculating an employee's weekly pay, (b) deciding which of three suppliers to use using cost data and past reliability, (c) deciding the company's strategy for the next ten years.Show worked solution →
(a) Structured: fixed rules (hours times rate, tax tables) can be fully automated.
(b) Semi-structured: the comparison of costs and reliability follows set steps, but choosing between close options needs judgement.
(c) Unstructured: no set procedure; it relies on judgement about an uncertain future.
Marking guide: 1 mark each.
core4 marksDescribe how a hospital could use a decision support system to plan staff rosters for its emergency department.Show worked solution →
The DSS draws on data: years of patient arrival records, current bookings, staff availability, public holidays and local events.
Its model base forecasts patient numbers for each hour of the coming weeks and runs what-if scenarios (for example, a flu outbreak increasing arrivals by 20%).
The user interface shows managers forecast demand against planned staff on a chart and flags shifts that are understaffed.
The manager uses this analysis plus judgement (staff wellbeing, skill mix) to finalise the roster, so the decision is semi-structured and supported rather than automated.
Marking guide: 1 mark each for data, models, interface and the human decision role.
exam6 marksUsing a bank as the enterprise, explain the three categories of decision-making within decision support systems, and evaluate the risks of treating a semi-structured decision as fully structured.Show worked solution →
- Structured (automated decisions)
- Blocking a card after three incorrect PIN attempts follows a fixed rule and is automated with no human involvement.
- Semi-structured (a specified, finite set of instructions plus judgement)
- A home loan application is scored against set criteria (income, deposit, credit history), but an officer reviews borderline cases and considers factors the rules do not capture, such as a recent job change to a higher-paid role.
- Unstructured (judgement required)
- Deciding whether to open branches in a new country involves many uncertain factors; a DSS provides market data and scenarios, but executives use judgement.
- Risks of over-automation
- If loan approval is treated as fully structured, the system applies rules rigidly: applicants with unusual but sound circumstances may be unfairly rejected, errors or bias in the rules and data are applied at scale without review, and nobody is accountable for explaining decisions. It may also breach expectations of fairness and transparency.
- Evaluation
- Automation increases speed and consistency, but for semi-structured decisions that affect people's lives, a human should review edge cases and the system should explain its recommendations.
Marking guide: 1 mark per category with a bank example (3 marks), 2 marks for risks, 1 mark for a justified evaluation.