Mathematics
Riemann vs. Lebesgue Integral
Quick fact
The Lebesgue integral can integrate functions that the Riemann integral cannot, and it makes the limit theorems (like the monotone convergence theorem) work more cleanly, which is why it's the standard in modern analysis and probability.
Why this is interesting
You've probably computed areas under curves with the Riemann integral, but when functions get wild, it fails. What if there were a different way to integrate that could handle almost any function you can imagine?
Read the full explanation
Understanding Riemann vs. Lebesgue Integral
Think of the Riemann integral as slicing the x-axis into vertical strips and adding up their areas. The Lebesgue integral, instead, slices the y-axis: it groups points where the function has similar values, weights each group by the "size" (measure) of that set of points, and sums those weighted values. More formally, the Riemann integral approximates a function by partitioning its domain into intervals and using rectangles. The Lebesgue integral partitions the range into intervals and looks at the preimage of each interval, measuring how much of the domain maps into that range interval. This shift allows the Lebesgue integral to handle functions with many discontinuities or wild oscillations, because it only cares about the measure of sets where the function misbehaves, not the pointwise details.
A deeper explanation
The core difference lies in two perspectives: partitioning the domain (Riemann) versus partitioning the range (Lebesgue). The Riemann integral is defined via upper and lower Darboux sums, which fail for functions that oscillate too much (e.g., the Dirichlet function). The Lebesgue integral, built on measure theory, assigns a measure to sets of points, and the integral is defined as the supremum of integrals of simple functions dominating the given function. This allows the Lebesgue integral to integrate functions that are not Riemann integrable, and it satisfies powerful convergence theorems under mild conditions. The Lebesgue integral is essential in probability (where expectation is an integral over a sample space), in the study of complete normed function spaces (Lp spaces), and it forms the basis for distribution theory and Sobolev spaces used in PDEs.