Energy management in buildings requires optimization and it is much better to optimize the entire system than to optimize each subsystem alone. That is to say, sub-optimization could lead to misleading results as this approach ignores the "interaction effects" among the subsystems.
It would be easier then to carry out a global sensitivity analysis of the model associated to the considered physical system in order to identify the most influential parameters while observing a key output parameter of this model.
Thanks all for your usefull information. I just finished my PhD in the field of energy , buildings and renewable energy systems and thinking of working in the area of whole system modelling instead of sub-system modelling. I am going to extend my knowledge to find a research position around this area.