A method for generating all efficient solutions of 0-1 multi-objective linear programming problem (Q2571939)

The following (0-1) multi-objective linear programming problem (further abbreviated as (0-1)-MOLP) is considered: \[ \text{Max}(C_1 W,\dots, C_sW)\quad\text{subject to }W= (w_1,\dots, w_n)\in X\subset\mathbb{R}^n, \] where \(X= \{W\mid AW\leq b,\,w_j\in \{0,1\}\) for \(j= 1,\dots, n\}\), \(c_i\in\mathbb{R}^n\) for \(i= 1,\dots, s\), \(b\in \mathbb{R}^m\), and \(A\) is a \((m,n)\)-matrix. A one-stage algorithm, which determines at least one efficient solution in each iteration is presented. This algorithm makes possible to find all efficient solutions of the given (0-1)-MOLP, without generating all feasible solutions. The algorithm is demonstrated on a small numerical example (3 objective functions, 5 variables, 3 linear constraints). In each iteration some new constraint has to be added so that the problem becomes in each iteration larger, which is a disadvantage of the proposed method.