Optimal choice problem?

Question

Programming

There is a problem in which it is necessary to find a globally optimal solution. The fact is that the choice made in each of the subtasks is not consistent with the others, so the greedy algorithm is categorically not suitable. Yes, it sounds too abstract, but: what are the mathematical ways to solve the problem?

Answer 1

Abstractly - google for the phrase "optimization methods"
A little more specifically - set the optimality criteria and evaluation methods for a particular parameter. You can also set the weight of this parameter. Formalize it in any convenient language and use the scientific poke method to find the optimal combination of these parameters.

Answer 2

In general, it's a total bust.
Without exhaustive enumeration, the global optimum may not be found. One of the quick algorithms is genetic: first, a random selection is made, then several best options are selected at each step (group A) and three groups pass to the next generation: group A, children of group A (random "genes" of a pair from group A are crossed), as well as mutations from group A (a few random "genes" change).