I want to compute optimal solution to a problem. The problem that I am addressing cannot be divided into independent subproblems as the cost function shows strong dependence between subproblems.
It seems to me that the wording of your question is not clear enough, it is the dependence on structural system level or it is an optimization problem with multiple criteria. Every time I allow myself to suggest some references attached.
If your problem is unconstrained N. Z. Shor's method of generalized gradient descent with space dilation in the direction of the difference of two successive subgradients (the so-called r-algorithm) will be useful.
You mention "subproblems" - how are they defined in terms of the "full" problem? Is there some problem structure that can be utilized, such as a separability of some subvectors in the objective function? As it is, your explanation about the problem properties is rather limited.