Designing automated smart systems and assessing decision support output: HSC Enterprise Computing Intelligent Systems
“Verify the sources of data used in a decision support system; design and model an automated smart system using a range of inputs and outputs; implement automated processing using software; assess the output produced by a decision support system, including graphing to detail the success of decisions, and comparing the proposed versus the actual outputs”
Verify a DSS's data sources for authority, currency, completeness, accuracy, bias and licensing. Design automated smart systems as inputs, processing rules or models, outputs and feedback, with safety defaults, then implement them in software such as scripts, macros or microcontroller programs. Assess DSS output by graphing and comparing proposed against actual outcomes and refining the model.
Jump to a section
What this dot point is asking
These practical content points ask you to check the data behind a decision support system, design and build an automated smart system, and judge how good a DSS's recommendations were by graphing and comparing them with what actually happened.
The answer
Verifying the sources of data in a DSS
Before trusting a DSS, verify its data:
- Authority: who produced it (a government agency, the enterprise's own systems, an unknown website)?
- Currency: how recent is it, and how often is it updated?
- Method and completeness: how was it collected, and are there gaps?
- Accuracy and consistency: does it agree with other independent sources and with known facts?
- Bias: does the sample or the collector have a bias?
- Licensing: is the enterprise allowed to use it?
Designing and modelling an automated smart system
A smart system follows input, process, output, usually with feedback:
- Inputs: sensors (temperature, motion, light, moisture, location), user settings, external data (weather forecasts, prices).
- Processing: rules with thresholds, schedules, or models (including machine learning) that decide what to do.
- Outputs: actuators (motors, valves, relays), displays, alerts, data logs.
- Feedback: outputs change the environment, which the sensors measure again.
Model it with a flowchart or state diagram, a list of rules, and a table of inputs and outputs, before building. Include safety defaults, manual override and handling of faulty sensor readings.
Implementing automated processing using software
Automation can be implemented with:
- Spreadsheet macros and scripts that import data, recalculate and produce reports.
- Workflow automation tools that trigger actions when events occur (a form submission creates a task and sends an email).
- Microcontroller programs that read sensors and control actuators.
- Scheduled database queries or scripts that process data overnight.
Assessing DSS output
- Graphing proposed (recommended or forecast) values against actual outcomes over time shows how closely they match and where they diverge.
- Comparing proposed versus actual outputs numerically: differences, percentage error, and whether decisions based on the output achieved their goals (less waste, fewer stockouts).
- Look for patterns in errors (always too high on Mondays) that point to missing inputs or wrong assumptions, then refine the model and data sources.
A council's DSS recommends how many street sweepers to roster each week based on forecast leaf fall.
- Verify data: leaf-fall estimates come from Bureau of Meteorology wind forecasts and three years of the council's own collection records; both are authoritative and current.
- Graph: proposed sweeper hours and actual hours needed over 12 weeks on one line chart.
- Compare: the DSS was within 5% for 10 weeks but under-predicted by 30% in two stormy weeks.
- Assess: accurate in normal conditions; add a rule to increase recommendations when strong winds are forecast.
- Designing without safety logic
- Include fail-safe states and manual override.
- Assessing output only by whether it "looks right"
- Graph and quantify differences.
- Assuming data is correct because it is in a system
- Verify sources.
Practice questions
Original practice questions graded from foundation to exam level, each with a full worked solution. Try them before revealing the solution.
foundation3 marksIdentify the inputs, processing and outputs of an automated smart lighting system for a school hallway.Show worked solution →
- Inputs: motion sensor, light-level sensor, clock (school hours).
- Processing: IF motion detected AND light level is low AND within school hours THEN turn lights on for 10 minutes; ELSE keep off.
- Outputs: relays switching the lights; a log of usage sent to the facilities dashboard.
Marking guide: 1 mark each.
core4 marksA café's DSS forecast the number of muffins to bake each day. Over five days it proposed 40, 45, 50, 42 and 60, and the café sold 38, 47, 35, 41 and 58. Assess the DSS output.Show worked solution →
- Differences (proposed minus actual)
- +2, -2, +15, +1, +2. Four days were within 2 muffins, which is accurate. Day 3 overestimated by 15.
- Graph
- plotting proposed and actual on a line chart shows the two lines tracking closely except for Day 3, making the outlier obvious.
- Assessment
- the DSS is generally reliable, but the Day 3 error needs investigating (was it an unexpected event such as heavy rain or a nearby closure?). If the cause is predictable (weather), adding weather forecast data to the model would improve accuracy.
Marking guide: 1 mark for calculating differences, 1 mark for describing a suitable graph, 1 mark for judging overall accuracy, 1 mark for an improvement.
exam6 marksDesign and model an automated smart system for a small greenhouse. Include a range of inputs and outputs, the processing logic, how you would implement it in software, and how you would test it.Show worked solution →
Inputs: temperature sensor, humidity sensor, soil moisture sensors, light sensor, a manual override switch.
Outputs: vent motor (actuator), fan, drip irrigation valve, grow lights, and notifications to the owner's phone.
Processing model:
- IF temperature > 30 degrees THEN open vents and turn on fan.
- IF soil moisture < 25% AND time between 6 am and 8 am THEN open irrigation valve for 10 minutes.
- IF light level < threshold AND time between 6 am and 6 pm THEN grow lights on.
- IF any sensor reads outside plausible range THEN send an alert and keep last safe state.
- Manual override takes priority.
Implementation: program a microcontroller (for example in Python or C) to read sensors every minute, apply the rules, drive the outputs through relays, and log readings to a cloud spreadsheet for a dashboard.
Testing: test each rule with simulated sensor values (normal, boundary such as exactly 30 degrees, and faulty readings), then run a week-long trial comparing the intended conditions with actual logged conditions and adjusting thresholds.
Marking guide: 1 mark for inputs, 1 mark for outputs, 2 marks for processing logic including safety or override, 1 mark for software implementation, 1 mark for testing.