Criteria for efficiency in vector optimization
We consider unconstrained finite dimensional multi-criteria optimization problems, where the objective functions are continuously differentiable. Motivated by previous work of Brosowski and da Silva (1994), we suggest a number of tests (TEST 1–4) to detect, whether a certain point is a locally (weakly) efficient solution for the underlying vector optimization problem or not. Our aim is to show: the points, at which none of the TESTs 1–4 can be applied, form a nowhere dense set in the state space. TESTs 1 and 2 are exactly those proposed by Brosowski and da Silva. TEST 3 deals with a local constant behavior of at least one of the objective functions. TEST 4 includes some conditions on the gradients of objective functions satisfied locally around the point of interest. It is formulated as a Conjecture. It is proven under additional assumptions on the objective functions, such as linear independence of the gradients, convexity or directional monotonicity.
KeywordsVector optimization Efficiency criteria Density
Mathematics Subject Classification (2000)49D39 65F99 15A39
Unable to display preview. Download preview PDF.
- Demjanov VF, Malozemov VN (1974) Introduction to Minimax, translated from Russian by Louvish D. Wiley, New YorkGoogle Scholar