Welcome to Computational Methods!

Hi there! Ever wondered how software engineers tackle massive, complex problems like mapping the entire world for GPS or creating a computer that can beat a chess champion? They don't just start typing code immediately. Instead, they use a set of mental tools called Computational Methods.

In this chapter, you’ll learn how to look at a problem like a computer scientist. We will explore how to break down big tasks, find shortcuts to "good enough" answers, and model systems before they even exist. Don't worry if some of these terms sound scary at first—we'll break them down step-by-step with plenty of everyday examples!

1. What Makes a Problem Solvable?

Before we can solve a problem with a computer, we have to decide if it is amenable to a computational approach. This simply means: can a computer actually handle it?

Features that make a problem solvable by a computer include:

  • Clear Inputs and Outputs: We know exactly what data goes in and what we expect to get out.
  • Defined Rules: The problem follows a logical set of steps (an algorithm).
  • Finite Time: The computer must be able to reach an answer in a reasonable amount of time.

Example: Calculating the shortest route between two cities is solvable. Predicting exactly what you will want to eat for dinner in ten years' time? Not so much—there are too many unpredictable variables!

2. The Core "Thinking" Methods

These three techniques are the foundation of problem-solving. You might have seen these in the "Computational Thinking" section, but here we look at them as practical methods for building software.

Problem Recognition

This is about looking at a new problem and asking, "Wait, have I seen something like this before?" Computers are great at solving patterns. If you recognize that a new problem is just a version of an old one, you can reuse existing algorithms.

Problem Decomposition

This is the process of breaking a complex problem down into smaller, more manageable parts called sub-problems.
Analogy: If you were tasked with "Organising a Music Festival," you’d be overwhelmed. But if you decompose it into "Book the bands," "Hire security," and "Sell tickets," the task becomes much easier to manage.

Use of Abstraction

Abstraction is the process of removing unnecessary details so you can focus on the important parts.
Analogy: A London Underground map is a perfect abstraction. It doesn't show the exact curves of the tracks or the depth of the tunnels; it only shows the stations and the connections, because that’s all a passenger needs to know.

Quick Review: The Big Three

Recognition: Finding patterns.
Decomposition: Breaking it down.
Abstraction: Hiding the "extra" stuff.

3. Divide and Conquer

This is a specific, powerful method of problem-solving. It works by:
1. Dividing a problem into smaller versions of itself.
2. Solving the smallest version (the "base case").
3. Conquering (combining) the solutions back together.

Many famous algorithms, like Merge Sort or Binary Search, use this method. It is often much faster than trying to solve a huge problem all at once.

4. The Programmer's Toolkit

Sometimes, basic logic isn't enough. Here are the specialised methods you need to know for the OCR syllabus:

Backtracking

This is a "trial and error" method. You start down one path to find a solution; if you hit a dead end, you backtrack to the last successful point and try a different path.

Analogy: Solving a maze. If you hit a wall, you don't stay there; you go back to the last fork in the path and try the other way.

Data Mining

This involves searching through massive sets of data (Big Data) to find hidden patterns or relationships that aren't obvious to humans.
Did you know? Supermarkets use data mining on loyalty card data to realize that people who buy nappies often buy beer at the same time, so they place them near each other in the aisle!

Heuristics

Sometimes, finding the perfect answer takes too long. A heuristic is a "rule of thumb" or a "best guess" that finds a good enough solution quickly.
Example: The "A* Algorithm" uses heuristics to find a path in a video game. It doesn't check every single atom on the map; it just moves in the general direction of the goal.

Performance Modelling

Before building a massive system (like a new banking network), we create mathematical models to see how it will perform under stress. This helps avoid crashes before the "real" version is ever built.

Pipelining

This is a method to improve efficiency by starting the next task before the current one is finished.
Analogy: In a car factory, workers don't wait for one car to be completely finished before starting the next. While one car is getting its doors, the next one is already getting its engine.
Key Takeaway: Pipelining is about overlapping tasks to save time.

Visualisation

This is the process of turning complex data into a visual format (like a graph or a 3D model). This makes it much easier for humans to spot trends, outliers, or errors that would be invisible in a giant list of numbers.

5. Summary Checklist

Before you move on, make sure you can explain these to a friend:

  • Can I explain why a problem might or might not be solvable? (Rules, inputs, outputs).
  • Do I know the difference between decomposition and abstraction? (Breaking down vs. hiding detail).
  • Can I explain "Divide and Conquer"? (Splitting the task into smaller identical tasks).
  • Do I know my toolkit? (Backtracking, Data Mining, Heuristics, Performance Modelling, Pipelining, Visualisation).

Common Mistake to Avoid: Don't confuse Pipelining with Concurrent Processing. Pipelining is about overlapping stages of different tasks in a sequence. Concurrent processing is about doing multiple tasks at the exact same time.

Don't worry if this seems like a lot of definitions! Just remember that these are all just different ways of being "lazy" in a smart way—finding the most efficient path to an answer!