When my opponent has white and plays 1. d4, I play the Queen's Gambit Declined: 1...d5 2. c4 e6. My problem is that sometimes my opponent plays c5 before I can prepare and play ...c5 myself. I have had trouble playing against this system because of the space disadvantage for Black. Is there a good way to prevent White from playing c5 or taking advantage of it if he does?

play the tarrasach if you want to.  That plays c5 on move 3, after e6. 

How early is White playing c5?  Doesnt it reduce central tension if too early and overextend a pawn chain?  Is it one pawn move too many when development could be continued?  Can you build up and hit out with e5 in response?  I know I am just asking questions in response but I am not sure at what stage you mean.

Playing b6 followed by a5 (if white plays b4) will solve white's space advantage "problem".

what he said

@MetaKnight: Thank you for that suggestion. I will try the Tarrasch and see how it works.

@HereIsPlenty: Normally it's around the 7th move or so. In the last game I played where White played c5, White had already played a3 and was able to consolidate with b4. Here is how it started:

I responded with c6 and later played b6, where my opponent played b4. I did eventually play e5 and won the game when my opponent dropped a rook on a discovered attack, but I feel like I handled the opening poorly

c6, probably followed by e5 seems like a pretty good approach to me.

In that or in similar cases, esp when it's early, you attack the pawn chain with moves like e5, b6 and if white plays b4 to support it play a5.

Often with piece pressure e.g. Qb6, c7 or f6.  Knights can often help with Nbd2 or Nf6-e4.  This is only when white insists on trying to keep this space advantage.

If it does not win a pawn, it usually leaves white with poor pawn structure or behind in development.  c5 isn't always bad for white, but this early in the game it's too far up the board to support when black attacks it.

Well, weak openings will tend to have a very small draw percentage.  Strong openings by world class players will obviously have the highest drawing percentage.

lol @ all the low rated players concerning themselves with how drawish an opening is. It doesn't honestly matter under a certain elo point since there are enough errors in the middlegame and endgame to allow for winning chances regardless of how drawish your opening was. More important is knowing the structure and general plans of the opening you're playing rather than knowing how well is fairs in master games.

On topic, @OP: don't worry about c5 for white. A space advantage is only dangerous if it stifles play or development. In given cases, black has no trouble developing or gaining play through b6 and e5 breaks. c5 by white is usually a bad choice by white.

Maybe Slav-complex is played most by masters in the last ten years, but in most multi-million game databases they are a small fraction of the QGD games.  And only a tiny % of Slav games are the Exchange Variation, and by no means are all of them drawn.

The claim the whole QGD is drawish is just laughable.

One could even argue that the Exchange Slav is far less drawish than percentages claim, as it's often used to draw instead of play chess. One of the reasons the percentages are so high is that people who want to draw as white just trade everything on the c-file and agree to a draw by move 20, while people who are playing for a win in the opening can usually find a slight but enduring advantage which can often be pressed into a win.  Also note that the Exchange Variation is commonly used for prearranged draws, as no one will think anything of a draw in it.  One could say that a main reason it is so drawish is due to its reputation building on itself.

EDIT:

Amusingly enough, a lot of my wins in the Exchange have come against players higher rated than I am who thought I was playing for a draw and treated my opening choice with disgust.  Anyway though, the main point of what I'm saying is that most openings are drawish if you don't make a serious effort to win, and that most openings have winning chances if you do.

When White plays an early c4-c5 in the QGD he is releasing the central tension AND foregoing the chance to emerge with a pawn majority in the center after an eventual ...d5xc4, so he is putting a lot of faith in the extra Queenside space and cramping effect of c4-c5.

One problem Black may find is his Bc8 trapped behind his own pawns fixed on light squares, so the most direct solution is to play ...b6 and if b2b4, as suggested by proknight98 on the 1st page.  Black can then trade off his problem piece at a6, and then can trade Rooks on the a-file if he wishes.

paulgottleib's suggestion of aiming for ...e5 is strategically sound, but it is easiest to play ...c6, ...b6, ...a5 etc. first.  Then the Queenside will be either neutralized or soon blocked, and Black can turn his attention to the counterthrust in the center - which is of course the indicated response to a flank demonstration by White anyway.

Bottom line is that Black should have nothing to fear from an early c4-c5 as long as he is careful.  White gives up more options than he gains with the push.

And as far as the actual game with 3 a3, that is quite a nothing move and NM iFrancisco points out that Black can respond with 4...c5! and avoid the problem and get a good game earlier.

The most common instances of an early White c4-c5 that is viable would be the [D15] Chebanenko Slav 2 c4 c6  3 Nf3 Nf6  4 Nc3 a6  5 c5, or the [D37] 2 c4 e6  3 Nc3 Be7  4 Nf3 Nf6  5 Bf4 0-0  6 e3 Nbd7  7 c5.  As with many new ideas, c4-c5 scored very well at first in these lines, but Black figured it out soon enough.