On the effects of combining objectives in multi-objective optimization (Q494690): Difference between revisions

The authors consider the multiobjective problem (P) of minimizing a vector function \(f: \mathbb R^n\to \mathbb R^m\) on a nonempty set \(S\) with respect to a convex cone \(K\subset \mathbb R^m\). They establish a relationship between the optimal solution set of (P) and the solution set of the problem of minimizing \(Af\) over \(S\) with respect to \(C\) where \(A\) is a \(p\times m\) matrix and \(C\) is a convex cone in \(\mathbb R^p\) such that \(K=\{y\in \mathbb R^m: Ay\in C\}\). They find also some conditions under which a given proper efficent solution \(\bar{x}\) of (P) has a unique scalarizing vector \(w\) such that \(\bar{x}\) minimizes \(w^Tf(x)\) over \(S\). This later result gives an anwser to the question of how to change \(K\) so that \(\bar{x}\) is no longer an optimal solution of the problem (P).