How long has dark matter been around? I've read that dark matter's existence is theorized because galaxies spin/clump in a way that would require much more mass than there appears to be. Also that there is gravitational lensing in places where there does not appear to be as much mass as would be required for the amount of lensing observed. So it is assumed there is some mass that we can't see, which is called dark matter.
Is there any evidence about how long dark matter has been around? Has the amount of dark matter been constant since the big bang or is it increasing? If it is increasing, is regular matter decreasing at an equal rate?
 A: It depends upon which dark matter theory you follow: some dark matter theories predict that dark matter particles have rest masses in TeVs (super-WIMP), while others predict masses in keVs (certain Warm Dark Matter-WDM-candidates).
The expected rest mass of WIMPs is in GeV-TeVs, so we'd need an extremely high temperature and density for their formation, which could only have been achieved in the early universe. It's unlikely that appropriate conditions are abundant now, even during supernovae (see this physics SE post: When and where was WIMP dark matter formed?). But clearly it'd be easier to achieve the most basic condition: energy availability, needed for the formation of lighter particles (there may be other factors to consider as well). For example, if sterile neutrino theories (among other WDMs) are correct, there could conceivably be a lot of dark matter formation going on (provided there are no other caveats which are difficult to satisfy).
There are even theories about Late-Forming dark matter, which predict that dark matter formed after the epoch of Big Bang nucleosynthesis, but before the period of the decoupling of the microwave background.

For similar reasons, to answer the question "how constant have the levels of dark matter remained?" you'll need to consider the model you're using. Scientific communities are presently leaning towards the WIMP model, and hence the theories of dark matter annihilation are relevant to this discussion.
According to our definitions, WIMPs do not have electrical charge; WIMPs are their own anti-particles, so a process of 2 WIMPs annihilating to yield a fermion and its antiparticle (or even a gague boson pair of $W^+$ and $W^-$ or 2 $Z^0$ particles) can take place without violating quantum number conservation. Indeed, this is probably the process through which WIMPs formed in the early universe, with the abundant energy needed for the high mass.
One proposed theory about how this process is involved in the changing quantities of DM involves the process of DM annihilation affecting the evolution of a star, and is described in this paper on arxiv.

Salati, Pierre. “Dark Matter Annihilation in the Universe.” International Journal of Modern Physics: Conference Series, vol. 30, 2014, p. 1460256., doi:10.1142/s2010194514602567.

Paraphrased and simplified, it says that a dark matter halo may collapse into dense regions of WIMPs in the center of a galaxy. Stars nearby could attract some of these WIMPs; the process of annihilation of pairs of WIMPs subsequently takes place and the star evolves into a red giant.
A: 
How long has dark matter been around?

Short Answer
Almost forever.
Long Answer
Assuming a dark matter particle paradigm, according to a pre-print by Yang (2015) subsequently published in Physical Review D, the lower bound on the mean lifetime of dark matter particles is $3.57\times 10^{24}$ seconds. This is roughly $10^{17}$ years. By comparison the age of the universe is roughly $1.38 \times 10^{10}$ years.
This means that dark matter (if it exists) is at least as stable as anything other than a proton, which has an experimentally determined mean lifetime of at least $10^{34}$ years, or an electron, which is theoretically stable (just as the proton is in the Standard Model) and has an experimentally determined mean lifetime of at least $6.6×10^{28}$ years.
This means that all dark matter candidates that are not perfectly stable or at least metastable are ruled out. Decaying dark matter and dark matter with any significant annihilation cross section are inconsistent with observation, unless there is a mechanism that generates new dark matter in equilibrium with the amount annihilated.

Has the amount of dark matter been constant since the big bang or is
it increasing? If it is increasing, is regular matter decreasing at an
equal rate?

The LambdaCDM "Standard Model of Cosmology" assumes a constant amount of dark matter in the universe after the earliest moments of the universe (with the density of the dark matter in the universe decreasing in proportion to the spatial volume of the universe), just as the model does in the case of ordinary baryonic matter.
For purposes of this question, exactly how many moments after the Big Bang it takes for dark matter to emerge is pretty much irrelevant, as this number is much smaller (by a factor of many billions) than margin of error in our estimates of the age of the universe.
Caveat
Not all lines of evidence are consistent with this analysis, however. An article in the journal Nature, Bowman (March 2018), analyzing the "21 centimeter line" in the radio spectrum finds that:

[E]ither the primordial gas was much colder than expected or the
background radiation temperature was hotter than expected.
Astrophysical phenomena (such as radiation from stars and stellar
remnants) are unlikely to account for this discrepancy; of the
proposed extensions to the standard model of cosmology and particle
physics, only cooling of the gas as a result of interactions between
dark matter and baryons seems to explain the observed amplitude.

In other words, this evidence contradicts the LambdaCDM model, which assumes that dark matter is "almost collisionless" and hence could not cause massive cooling through interactions between dark matter and baryons. This evidence is consistent with an early universe (post-radiation era, hundreds of millions of years after the Big Bang) that has no dark matter, but that possibility throws a wrench into other aspects of the LambdaCDM model.
This contradiction, just recognized a few months ago, has not yet been adequately resolved.
