Definition
Problem recognition involves clearly defining a problem by understanding stakeholder needs and translating them into specific system requirements.
Key Aspects of Problem Recognition
Key Point
| Problem-recognition activity | Key questions |
|---|---|
| Analyse current solutions | What works now, what fails, what can improve? |
| Identify data requirements | What inputs/outputs/storage/scale are needed? |
| Stakeholder involvement | Are we solving the real user/business problem? |
Recognising a Problem From a Scenario
Problem recognition starts with working out, from the description given, exactly what has to be solved: what data goes in, what must come out, and whether a computer can produce it. It also means noticing when a new problem is a version of one that already has a known solution.
Worked Example
Problem Recognition: Delivery Driver Schedule
Once the problem has been recognised as a routing problem, existing algorithms such as Dijkstra's shortest path can be reused in place of a solution designed from nothing.
| What the scenario says | What is recognised |
|---|---|
| Each driver must visit a list of addresses using as little fuel as possible | A shortest-route problem, like the travelling salesman problem |
| The addresses are joined by roads of known length | The data can be modelled as a weighted graph |
| Some parcels must arrive before midday | A constraint that every route has to satisfy |
| The routes are needed within seconds each morning | An exact answer may take too long, so a heuristic is appropriate |