Computer Science I / Integration and Antiderivatives
Practice question · Multiple choice

Solving a linear system by computing the inverse is correct mathematics and discouraged in code. What are the two objections?

Hints
  1. Computing A⁻¹ means solving Ax = e for each column of the identity. How many systems is that?
  2. Every floating-point operation rounds. Which route performs more?
Show the answer

C. Cost and accuracy: it solves n systems where you needed one

Why

Roughly n³ against n³/3, plus more roundings on an ill-conditioned matrix. The formula x = A⁻¹b states what the solution is; it was never an instruction for computing it, which is why solve exists and calling inv first is the classic tell of someone translating algebra directly into code.

Read the lesson: Integration and Antiderivatives →

Practise Integration and Antiderivatives

The app has 7 more questions on this lesson, and keeps your place in the course. Computer Science I is free to start.

More questions on Integration and Antiderivatives