You are correct that ΔG
o for the aldolase reaction is +24 kJ/mol. However, remember that the ΔG
o refers to the free energy change under standard conditions (all reactants and products at 1M concentration). The true free energy change associated with a reaction depends on the relative concentrations of products and reactants: ΔG = ΔG
o + RT ln Q, where Q is the reaction quotient (in this case [DHAP][G3P]/[FBP]). Thus, a reaction with a positive ΔG
o can still occur spontaneously if you have a large excess of reactant and a small amount of product (i.e. Q << 1).
Indeed, in glycolysis, the aldolase reaction happens after a particularly favorable reaction (the phosphofructokinase reaction) which leads to the build up of FBP. In this way, ATP hydrolysis in the PFK reaction is being indirectly coupled to the aldolase reaction in order to help drive the reaction forward. Similarly, products get depleted by thermodynamically favorable reactions later on in glycolysis (in particular the 3-phosphoglycerate kinase reaction).
However, you are correct, that aldolase should be able to run in reverse. In cells undergoing gluconeogenesis, where the favorable PFK reaction is inhibited, aldolase does run in reverse, producing FBP from DHAP and G3P.
For more information see
http://onlinelibrary.wiley.com/doi/10.1016/0307-4412%2881%2990102-3/abstract