suppose that we can describe a problem with two formulas, i.e. we have:
f(a,b) and g(b,c) with unknown a,b,c
with
a=inv(c)
and we have such optimization algorithm as a solution:
-1: start with random a, name it old-a
-2: solve f(old-a,b), and find b
-3: put b in g(b,c), and find c
-4: now we have an approximation of a: new-a = inv(c)
-5: old-a = new-a, go to step 1, until we reached to an err such that err=(new-a - old-a)<epsilon
What is the name of this optimization?
I know it could be a multi-objective optimization, goal programming, a kind of linear programming, or some chain ... but I want a certain name of optimization problem or sub-class name, with an exact definition of it. I mean, if it is a sub-class of LP, then what is the sub-class name, and where is a reference book or paper?
I asked this question from several students, and they said that it is not a special algorithm, it is just a kind of evolutionary optimization.