We want to compare algorithms to decide which we prefer.
- 1.
- But there are many dimensions along which we might compare.
So we decide that what really matters is run time.
- 2.
- But run time depends on input. So we decide that what really
matters is run time as a function of input size.
- 3.
- But this can be hard to characterize. So we decide that what
really matters is worst-case run time as a function of input
size.
- 4.
- But one algorithm might be better on small inputs and the
other on large inputs. So we decide that what really matters is
worst-case run time as a function of input size for large
inputs.
- 5.
- But this can still be quite hard to determine precisely. So
we decide that what really matters is worst-case run time as a
function of input size for large inputs, ignoring constant
factors.
Next: Asymptotic Growth
Up: ASYMPTOTIC GROWTH
Previous: ASYMPTOTIC GROWTH