This paper investigates how recently endorsed methods in distributed optimization can be exploited in the energy-management problem for heavy trucks, with focus on improved control of ancillary systems. The justification for investigating this is because it may offer higher modularity and integrity to the development process of vehicle models. This paper assumes an indirect approach for solving the optimal control problem. For simplicity, a multi-level control hierarchy is also assumed, where prediction and integer decisions are planned at a higher level, while real-time decisions take place at a lower level. This allows the higher level to deliver an estimate of optimal costates to the lower level. Convex models of ancillary components allow the problem to be formulated as a second-order cone program at the lower level, which can be reformulated as an exchange problem suitable for distributed control. The exchange problem is solved using the alternating direction method of multipliers with proximal message passing, which returns a solution with the same fuel economy as that from dynamic programming. Convergence properties are briefly discussed, where the most notable conclusion is that warm start gives a significant improvement on the convergence rate, supporting practical feasibility of the distributed approach.