I want to express it in the form

$$\text{expr=}\int_{-\infty}^{\infty}C_1\exp\left(-t^2\right)f(t)dt$$, i.e., the integrand is a bounded smooth function $f(t)$ times $\exp(-t^2)$ and the domain of integration is the whole line.


Define the integrand with the measure as

integrand = c*Exp[-m*x^2/ω^m]*1/(Sqrt[2*π]*λ*ω)*Exp[-(Log[B] - μ)^2/(2*λ^2)] Dt[B, Constants -> {λ, m, x, μ, ω}];

where we use Dt[t] to represent the measure, which we will transform using the substitution:

subs = Assuming[{μ > 0, λ > 0, B > 0, t ∈ Reals}, First@Solve[t == (Log[B] - μ)/(Sqrt[2] λ), B, Reals]]
(* {B -> E^((2 t λ + Sqrt[2] μ)/Sqrt[2])} *)

The transformation is done via

transformedIntegrand = integrand /. subs /. Log[Exp[x_]] :> x // Simplify
(* (c E^(-t^2 + Sqrt[2] t λ + μ - m x^2 ω^-m) Dt[t, Constants -> {m, x, λ, μ, ω}])/(Sqrt[π] ω) *)

where we have simplified the expression by hand to some extent.

Finally, we integrate:

Integrate[transformedIntegrand /. HoldPattern[Dt[a__]] :> 1, {t, -∞, ∞}]
(* (c E^(λ^2/2 + μ - m x^2 ω^-m))/ω *)

(We have to set the measure back to 1, because obviously in Mathematica you don't specify the measure explicitly.)

  • $\begingroup$ Thanks for you answer. Sorry, I mistype my MMA code...Please check your answer...$B$ in fact should be $\omega$. There is no $B$ in the expression.. $\endgroup$ – Srestha Narayanan May 5 '16 at 7:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.