The luminosity of Trappist-1 is estimated to be $5.25\times 10^{-4}\ L_{\odot}$, but it has not always been like this.
The luminosity of a brown dwarf decreases with time and it this measured luminosity (along with the spectral type) that allows an estimate of the mass and a lower limit to the age using stellar evolutionary models.
If I look at the Baraffe et al. (2015) low-mass evolutionary models and look at the locus of luminosity versus time for a $0.08\ M_{\odot}$ star like Trappist-1, you can see that the current luminosity implies an age of $\sim 500$ million years. But if you go back in time, the star was more luminous and for that reason, planets that are currently in the habitable zone (said to be planets e,f,g) were not so in the past.
The details of a habitable zone (HZ) calculation can be complex, but basically the radius of the habitable zone scales as the square root of the luminosity. If planets d and h are not currently in the HZ then we can use these as a conservative definition of the HZ boundary.
From this (and using the published orbital radii of the planets), I can see that if the luminosity is increased by a factor of 9, then none of the planets b-g are in the HZ, it is larger than all of their orbits. Trappist-1 had a luminosity that was 9 times larger when it was younger than 27 million years old. On the other hand, if I want to move the HZ just outside the orbit of planet e (and simultaneously include planet h inside the HZ), then this would occur when Trappist-1 was at an age of 206 million years. As a final thought, you can see from this particular model that Trappist-1 may fade by a further factor of two as it gets older. This decreases the HZ radius by a factor of 1.41 and would mean that g (and possibly f) would fall outside the HZ, whilst d (and possibly c) would be brought into the HZ.
It should be noted however that: different models give slightly different results, these loci are mass dependent and the mass is not known, it is inferred from the same models using a temperature estimate (which is also uncertain). So whilst my qualitative conclusions about the past location of the HZ are likely to be correct (though the detailed age numbers are model-dependent), the future behaviour of the HZ is more uncertain because Trappist-1 may be slightly more massive than assumed and already reached its minimum luminosity.
The luminosity evolution of Trappist-1, assuming a mass of $0.08M_{\odot}$ and the models of Baraffe et al. (2015). The horizontal dashed line marks the best estimate of its current luminosity, for which planets e-g are said to be in the HZ. If we go back in time, the rightmost vertical dashed line marks the age below which the luminosity increases to the point that e becomes too hot to be habitable. Then further back, the leftmost vertical dashed line marks the point where all the currently known planets (b-h) become uninhabitable.
So the answer to your question is quite uncertain and depends critically on the age of Trappist-1 now and of course whether the planets were always at the orbital radii they are now$^{*}$. As you can see from the plot above (note the logarithmic scale on the x-axis), the luminosity evolution referred to above takes place early-on. If Trappist-1 could be as young as 500 million years, then life on planet e may only have been possible for 300 million years. However, if the star is slightly more massive and is 10 billion years old, then life has had 9.8 billion years to get going.
If you are talking about planet f, then it has had slightly longer ($\sim +100$ million years) inside the HZ, and planet g slightly longer ($\sim +70$ million years) again. Planet h will have spent comparatively little time (in the past), less than a few hundred million years, inside the HZ.
$^{*}$ The abstract of the discovery paper by Gillon et al. (2017) briefly discusses the possibility that the planets migrated inwards after formation, through a process of "disc-driven migration". If so, then this will not alter the discussion above. Disc around very low-mass stars may be longer-lived than those around higher mass stars, but have essentially dispersed after $\sim 10-20$ million years (Kennedy & Kenyon 2009; Dawson et al. 2013; Binks & Jeffries 2017), and the planetary configuration would have to be settled where it is now by the time the disc has gone.
