Is there a way to approximate $y=-W(-e^{-x-1})$ for large value of x? (W is the lambert's function)

72 Views Asked by At

I was working on a problem about carnot heat pump and came across this equation $y=-W(-e^{-x-1})$ and when i plotted it, it seems to me that for large value of x it started becoming linear so i was wondering if there is a way I could approximate this for large value of x.dotted line represents $y=x+5$ and the black line represent $y=-W(-e^{-x-1})$

I've tried some approximation already. I tried using the lagrange inversion theorem to expand the Lambert's function and after some approximation i ended up with $y=e^{-x-1}(e^{-x-1}+1)$, then after i've plotted it (dotted line is after the approximation) it seems like it's working for only small values of y.

1

There are 1 best solutions below

1
On BEST ANSWER

Chatzigeorgiou proved bounds on the exact expression you have in 2013: for $x>0$ $$1+\sqrt{2x}+\frac23x<-W_{-1}(-e^{-x-1})<1+\sqrt{2x}+x$$ Tighter upper bounds are possible if $x$ is known to be bounded; if $x<1$, $\frac34x$ can replace $x$ in the upper bound.

Another approximation is given in the first section of the paper. Let $a=0.3205$, then $$-W_{-1}(-e^{-x-1})\approx1+x+\frac2a\left(1-\frac1{1+a\sqrt{x/2}}\right)$$


Ioannis Chatzigeorgiou (2013), "Bounds on the Lambert function and their Application to the Outage Analysis of User Cooperation", IEEE Communications Letters 17 (8), pp. 1505–1508