Newtonpedia ES
Home › Mathematics › Henstock–Kurzweil integral
Mathematics · 1957

Henstock–Kurzweil integral

A method for calculating the area under a curve for very irregular functions, generalizing the Riemann integral.

Year
1957
People involved
Arnaud Denjoy
Section
Mathematics

History

The Henstock–Kurzweil integral, also called the gauge integral or generalized Riemann integral, is a way of computing the area under a curve by adding up a function's values multiplied by small intervals. It generalizes the Riemann integral and, in some situations, handles functions that the Lebesgue integral cannot. It matters because it widens the set of integrable functions while keeping the original idea simple.

The Riemann integral split an interval into pieces of fixed width and summed each height times its width. It worked well for reasonable functions but struggled with singularities, points where a function blows up or is undefined. In 1912 the French mathematician Arnaud Denjoy proposed a definition that could integrate functions such as f(x) = (1/x)·sin(1/x³), which has a singularity at 0 and is not Lebesgue integrable. The natural approach was to integrate while excluding a small neighbourhood of the problem point, then let that neighbourhood shrink to zero. To build a general theory, Denjoy used transfinite induction over the possible types of singularities, which made his definition quite complicated.

Other mathematicians reached different definitions. Nikolai Luzin worked with variations on absolute continuity, and Oskar Perron studied continuous major and minor functions. It took some time to realise that the Perron and Denjoy integrals were in fact identical. The decisive change came in 1957, when the Czech mathematician Jaroslav Kurzweil found a definition elegantly similar in spirit to Riemann's original, which he called the gauge integral. The gauge is a positive function that sets, point by point, how wide each interval of the partition may be; if it is constant, one recovers the Riemann integral. In 1961 Ralph Henstock independently introduced a similar integral while studying extensions of Ward's work on the Perron integral. Because of these two contributions, it is now known as the Henstock–Kurzweil integral.

In essence, the intervals may be narrower near problem points, as dictated by the gauge. A result called Cousin's theorem guarantees that a suitable partition always exists for any gauge, so the definition is never satisfied vacuously. This flexibility gives the integral useful properties: it is linear, and Hake's theorem lets improper integrals be treated as proper ones. Every Lebesgue integrable function is also integrable in this sense, and the converse holds when both the function and its absolute value are integrable, which is why the integral is often described as a non-absolutely convergent version of the Lebesgue integral. Moreover, integrating the derivative of a differentiable function recovers that function up to a constant, simplifying the second fundamental theorem of calculus. Its simplicity led some educators to propose it should replace the Riemann integral in introductory calculus courses.

Today it is used in advanced mathematical analysis, where it allows convergence theorems with weaker assumptions than usual. One example shows its reach: the function equal to 1 at irrational points and 0 at rational points on [0,1] is not Riemann integrable, yet its gauge integral equals 1.

Timeline

  • 1912Arnaud Denjoy proposes an integral definition able to handle singularities such as those in f(x) = (1/x)·sin(1/x³).
  • 1957Jaroslav Kurzweil finds a definition similar to Riemann's, which he calls the gauge integral.
  • 1961Ralph Henstock independently introduces a similar integral, and the shared name becomes established.

Trivia and anecdotes

A name with several fathers

The integral is also known as the narrow Denjoy, Luzin or Perron integral, though today it carries the names of Henstock and Kurzweil.

A proposal for teaching

Because of its simplicity, some educators have argued that it should replace the Riemann integral in introductory calculus courses.

People involved

Other milestones in Mathematics